AI_Assistant/pcbjam
Watch
1
0
Fork
You've already forked pcbjam
 0
Code Issues Pull requests Projects Releases 4 Packages Wiki Activity Actions 2
pcbjam/tests/screenshot-manifest.json

3946 lines
112 KiB
JSON
Raw Normal View History

feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
{
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"_note": "Baselines live in R2, content-addressed by sha256 — the local baseline-screenshots/ tree is a cache (`npm run screenshots:fetch`). Regenerated by `npm run screenshots:promote`; never edit hashes by hand. The name list is authoritative for removed-screenshot detection.",
"version": 2,
"storage": {
"bucket": "pcbjam-ci-screenshots",
"keyPrefix": "sha256/"
},
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
"screenshots": [
{
"name": "01-loading.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a8908f64d24597fc9af19df4bff8d91aac0b6a584a546698e1af73e9990b0c3d",
"bytes": 6017,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "03-after-load.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ad71660243c16719d9a35c79e6393896f12fd8481b386fffec690238a06b1c7d",
"bytes": 28969,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "04-controls-tab.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "47edc72ce3c4a919f68a7a874c0b98928991cf4065d33f8f2ae235601337d80d",
"bytes": 38561,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "05-text-input-tab.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "131465acd65d48a7f848b7ee7e43b2afd9c140ea521e52257a3e52e0b94a3856",
"bytes": 32894,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "06-text-input-typed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "fd1b96aabc2e65fd832221b59b1af309aefd23ae7ad4fd34a649c6c4ec6c5d73",
"bytes": 35416,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "07-drawing-tab.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "44287d1340094654bd1d7c643edf0b90e709923d15b5cd18c5898647ee286519",
"bytes": 37636,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "08-drawing-done.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "36dd1ec39ae8a1d04ef3db7be866d1f1d190c1749af2e32e73caa0473be0fb54",
"bytes": 37547,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "09-lists-tab.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "42be94987a02de93737ac151b787448e051332a9c8bb47ace48346a6d7e6a0fb",
"bytes": 42424,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "10-lists-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "42be94987a02de93737ac151b787448e051332a9c8bb47ace48346a6d7e6a0fb",
"bytes": 42424,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "10b-choice-selected.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9086cf1da7128b832d68c2a2108264de44b74d7b85e79baa855f647c49a9ddfe",
"bytes": 42806,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "11-file-menu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7bc0bae8efd36bb3b20ee652adf1e8c851df67b3d7a695da64c8f9b8b3135dd6",
"bytes": 42434,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "12-help-menu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9086cf1da7128b832d68c2a2108264de44b74d7b85e79baa855f647c49a9ddfe",
"bytes": 42806,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "13-final.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "464f2025a79276fd74bbaf450f8b5f81ae00400d901f3940f2db232acbb33d08",
"bytes": 43092,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
{
"name": "3d-viewer-00-board-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2bb785970e78ac20dafe353a4c628161d6406f517045f9e7674bed75a466dca8",
"bytes": 252106,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "3d-viewer-models-pic_programmer-render.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e4e9a46c2a1f0360e9d36bccfdd4fe68550fa5673ce0012ed05c613df7a20d9f",
"bytes": 72189,
"width": 1084,
"height": 623
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "3d-viewer-models-pic_programmer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "251d1161cd3db748047df90c2bb434e91eb66cd68765183af35d7111b5a4ce36",
"bytes": 115779,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "3d-viewer-pic_programmer-render.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c99b4d6daf7f030d7d87ef004b3e19f13c2b52aaa1f1d32ddfe91e73915f0766",
"bytes": 73632,
"width": 1084,
"height": 623
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "3d-viewer-pic_programmer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7609604bac2b08f19bbba6f0ac162a17f537acb1c550eac04fa8d55210d15045",
"bytes": 115635,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "3d-viewer-titlebar.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "f472573ddb17ec190d80e7e13eb07e052faed7ef2c665ac2c13832c71795d810",
"bytes": 115497,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
{
"name": "appearance-00-layers.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "appearance-01-objects.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "64fc1685bba9b1c987fef78871498b3206ed709bb8a1991bc2c7a8576faf100b",
"bytes": 104485,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "appearance-02-nets.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "6867a91a6637a865457a6b370761792ac7c53267ffc6f048be5579ab43e7c66f",
"bytes": 84489,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "appearance-03-layers-again.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "601393abb3624741be75e36723abcd55fdd752767009bed2e278a6a294149f5e",
"bytes": 105568,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "appearance-10-layers-scrolled.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b6e6740f8b9a7cd6de6f4827c79b4b6d23d7325cf11a7d28dbd7b53e14d84824",
"bytes": 105299,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "appearance-11-layers-scrolled-back.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "appearance-20-objects-scrolled.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "77848309996306c017c4bd727e4d565912f2a41d71c4423e2f58560789959983",
"bytes": 108841,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "dafef7fbc0cdcd31bdc386e9128856b3cd57bf526870137f2dca7666cfbd4b30",
"bytes": 28703,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-02-panels.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "dafef7fbc0cdcd31bdc386e9128856b3cd57bf526870137f2dca7666cfbd4b30",
"bytes": 28703,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-03-close-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "dafef7fbc0cdcd31bdc386e9128856b3cd57bf526870137f2dca7666cfbd4b30",
"bytes": 28703,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-04-dragged.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0f08df98b9a49f9015f69db017642a8805b4fc095848a4f5e97e92bce033441f",
"bytes": 28711,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-05-multi-panel.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "685f9930941f2db9ddf75c2ab0c93199196754cdc7c6efd3d8ef7c69e9720c56",
"bytes": 28837,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-resize-01-hover.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "dafef7fbc0cdcd31bdc386e9128856b3cd57bf526870137f2dca7666cfbd4b30",
"bytes": 28703,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-resize-02-mid-drag-live.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1f440434987e994552bb9280b2c6e9a17ac9ed20a6e46873751bcdf56d331e13",
"bytes": 28750,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-resize-03-mid-drag.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "972aaa93464b1772438cc8640f02ad113c4fb90cc500d6cba1684974614bd093",
"bytes": 28841,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "aui-resize-04-after.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8650595db7868c5544dfa2434bf845c0188b935fa014527a00a44dace268124b",
"bytes": 28843,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "auinotebook-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "80c2bc70925f46be1630c490a1a3e3c735d4809aec420909f5a058510313b605",
"bytes": 44737,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "auinotebook-02-tab-switch.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0040913c308ec96ba64433a7d16493155852a2c019eee50e8b690c65d547b89e",
"bytes": 45173,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "auinotebook-03-add-tab.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "5f1cdc8401251589570794310ef88065a88144a0be729479faf93e1c576affbd",
"bytes": 38664,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "auinotebook-04-remove-tab.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "114b86c5d55c04170dd8ff2de64f082099cc166f49984366e055b3d53dc4a51d",
"bytes": 46391,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "auinotebook-05-tab-style.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ad5809808d5ecbea4fc86f1049523ce55eb84f547283d65fc826b4d90b384c50",
"bytes": 46574,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "bitmapbuttons-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "fdd47d8f19db4a3ef9ebddeb054be406dd6c24203cf78a02e67e4318c031be0e",
"bytes": 35720,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "bitmapbuttons-02-toolbar-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "aac4d3e9e7d08f47c47fd9bc3ce08a65a32d023b1e7d4f984a6507345da4ab43",
"bytes": 38468,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "bitmapbuttons-03-toggle.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "72e2e587c4bcd525ac3bb4727e1f729187ef156c346e0d24190ac5dfd8cf5620",
"bytes": 37136,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "bitmapbuttons-04-multi-toggle.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "6a09c1aa935d996fc3ee111f228274238cfaf84bb2d1cc19b528f445d128c4a8",
"bytes": 38601,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "bitmapbuttons-05-enable-toggle.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0d985c3b70c3444e2374d6ef0152ba93f4b297933b8600153bc20db4bdb69129",
"bytes": 37269,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "bitmapbuttons-06-shapes.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ef497f0b9a1c998d6f9eecfdc1ac9e7c37ed88d66e2f31f17ce5e4c31c582866",
"bytes": 40257,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "bitmapbuttons-07-artprovider.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "914163b441cdf861b82c41416ea5936d88816736ae00f4f33b0bd49438e92107",
"bytes": 38059,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "bitmask.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b377296d2ed20c6f7bf9959a9d5105bfed8b0b84822934c80039ea37178810e8",
"bytes": 52974,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "calculator-before-switch.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "aa488ec1f9167980667f9d8a05b9bc52f972ae7585a3c724cd1c57177e8012d5",
"bytes": 68846,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "calculator-color-code.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "195b98722464d71e28edb528a8645c8cd2a3f74932c44b9d50f332f75d6e8803",
"bytes": 71587,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "calculator-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "aa488ec1f9167980667f9d8a05b9bc52f972ae7585a3c724cd1c57177e8012d5",
"bytes": 68846,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "calendar-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "96dcfca53f8a31717ecdf2701d6d6f4442b56d8d15a536a47e158f5d11ba736b",
"bytes": 41850,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "calendar-02-select-date.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "daccdc5e2994d4f5774efc96a232d30328c916ccd73e6fe08121f3165fd4d5a9",
"bytes": 45185,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "calendar-03-next-month.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c9ac7a97caa2fca1b51f61198fffb025a932e557a4fab49674d3b309418fe383",
"bytes": 44001,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "calendar-04-prev-month.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ec11f6fad364e011abad25e8d922864929624586b919227a88e0c796bb6344a8",
"bytes": 43534,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "calendar-05-today.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8c5b486bb401a293925aeb46d7abe3c829fc5bd008673154df245501ff0a289f",
"bytes": 45149,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "clipboard-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "3aa56046961871848dc51c8516ea67a6d3e2159449bbcdc9de55f42dcb70ddec",
"bytes": 31734,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "clipboard-02-copy-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "fd2491c104397c5055ec216bd3023c6ad27389fcc8ce8ba870a6f838e10d86bb",
"bytes": 34601,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "clipboard-03-paste-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bd31e04494bf4bbc9b788fb88f74e2694e42138740b07e5b0a3a5563d3ac8987",
"bytes": 40429,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "clipboard-04-check-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d4904392e79d3b840a403a00ea01dea433be8f2542f03d588300c0e006d5af55",
"bytes": 39186,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "clipboard-05-clear-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "304576cbdff87140d27e2c796e853cf7646603d45fc975e2a68ff0b7210e1f4f",
"bytes": 39691,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "clipboard-06-full-flow.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2ad6763715076996901156240a946b864e14ca3356ed22bb27bea7974b152507",
"bytes": 49183,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "collapse-relayout-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ca5091fa4579e6272076ae18085b85c580cabfc0d0b51b2b3ca48a6f3f367f9d",
"bytes": 23744,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "collapse-relayout-02-expanded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e5a5330335a2be50a893aa0192d9d9f2390d99c1c74d1a51657fb1699f5f8a92",
"bytes": 28580,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "collapsible-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "782ab189948b2da5471c27776476f8bd3a6421908ee0831a301df81db5ff3442",
"bytes": 38001,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "collapsible-02-panes.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "782ab189948b2da5471c27776476f8bd3a6421908ee0831a301df81db5ff3442",
"bytes": 38001,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "collapsible-03-expanded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "782ab189948b2da5471c27776476f8bd3a6421908ee0831a301df81db5ff3442",
"bytes": 38001,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "collapsible-04-expand-all.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "da18f9825f03886035a67746a73e777859bf05f02e868c44c07866f2b8419f02",
"bytes": 49651,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "collapsible-05-collapse-all.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "60441feca9bdd64daf2f4b459579d171dd36301222a2a138b00dd9abfdd1e039",
"bytes": 32486,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "contextmenu-01-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "117fab7062f34b707c512caa0d6a8cbf70b28f5c29466fbe688be748c1761ce5",
"bytes": 15382,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "coroutine-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8ca29fadc686ce62fdc035f0968b185103cd983f516eadb4e5eb375e37ed7586",
"bytes": 125122,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "coroutine-02-summary.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8ca29fadc686ce62fdc035f0968b185103cd983f516eadb4e5eb375e37ed7586",
"bytes": 125122,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "coroutine-nested-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7421a1568c9ff5c51df6d138b4de8e582ad8616bc8ae8fef9ea6ed61955718c2",
"bytes": 92012,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "coroutine-nested-02-summary.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7421a1568c9ff5c51df6d138b4de8e582ad8616bc8ae8fef9ea6ed61955718c2",
"bytes": 92012,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b6bbe192c7942fac1729ded46107042989726fefbb2ec4138c4d989ff94d2068",
"bytes": 85061,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-02-list-populated.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b6bbe192c7942fac1729ded46107042989726fefbb2ec4138c4d989ff94d2068",
"bytes": 85061,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-03-list-selected.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "72f767ae63d9c9cd2b829d283504cbdebc237deaa561ab0823d31da3178d23c8",
"bytes": 87769,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-04-list-add.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e548e05eeca5652dccaa7e0055363c8b9cd607897e0f7e7dcb98d0fd7d97ec1b",
"bytes": 86848,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-05-tree-tab.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "24ffb5c04e9ac045b4dea7919f00ec7f2b58866673c7dcd7e87562964551c83f",
"bytes": 38002,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-06-tree-selected.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "3fa3937040259dbd1ff22d4d3598999b4b76d803b4f5a952c13b8e2bebd1140e",
"bytes": 39966,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-07-tree-expanded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1e14be6bf81cfab8738f3ce0a96838d1d1f5797d1bcfc0073ba3028c729e7e4d",
"bytes": 49344,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-08-tree-collapsed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "49dfbe819d6726e9de1c3fae34c2e27604f76ff692f7a47129954c0cf094a171",
"bytes": 37703,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-09-column-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2ada8bb4371279410ff0c6bd57a89f011f4ec13dbff2dce4f0ce6e7f6958b501",
"bytes": 87944,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataview-10-scrolled.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ab56605ea4f5262f2e1bd64d994572c990b284664e285ae6d40f7b3627d8bda",
"bytes": 83713,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataviewvirtual-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "3d34dd84b50c46676ee15890f209c78b818e6730e9bd6c6aa7a407a8827979cc",
"bytes": 127648,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataviewvirtual-02-large-dataset.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "3d34dd84b50c46676ee15890f209c78b818e6730e9bd6c6aa7a407a8827979cc",
"bytes": 127648,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataviewvirtual-03-scroll.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4f0d859f58c27fbbb17f565a4f87c884c2b23b342e7616feac39e9b92d922fcb",
"bytes": 138109,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataviewvirtual-04-selection.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "3d34dd84b50c46676ee15890f209c78b818e6730e9bd6c6aa7a407a8827979cc",
"bytes": 127648,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dataviewvirtual-05-zone-manager.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "3d34dd84b50c46676ee15890f209c78b818e6730e9bd6c6aa7a407a8827979cc",
"bytes": 127648,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialog-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c3a47f81680860f795523c85a00b44406df2c3f84e8cb5124465121795165d18",
"bytes": 29930,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialog-02-info-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "68945796e6c0d6d6fcd78230fa27216ea7822c9dd5956a489d7fba0603387941",
"bytes": 43319,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialog-03-yesno-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "359954015822573411ed52b4e2abec5c8cbd2815cc64c231e9c0019509a9a97c",
"bytes": 45221,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialog-04-error-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0819956debc59d567dcdc2930b32de7a524754db667f7ed81cd3b3ed52d8b25e",
"bytes": 42198,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialog-05-custom-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a285bbe8bd3c1dbd20e3e6b013508320c954835c67b04d62362321323724f615",
"bytes": 43477,
"width": 1280,
"height": 922
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-00-initial.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "3419d71893837b3c09157778b442bd8d4404757b26b44f904e1cfdd6284fcfb3",
"bytes": 28996,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-01-tab-selected.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d7cae66608c0e8d8e4445c6f46e17cedf45b58b3e0281d1a2c6a54e768d0510d",
"bytes": 42447,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-custom-closed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2bdb002868246fe38e480ca4fa5de8c887b10083cba919c8c4106803f444d56d",
"bytes": 58146,
"width": 1280,
"height": 922
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-custom-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "12a4b20144f153d1afae6aaf32663cc0780f2989654725e3090df8534b57871f",
"bytes": 58166,
"width": 1280,
"height": 922
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-full-flow.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b58a0ca7954518ba0f7efb20d80b1f668bdde32b0660b5298db98a20bc666cb8",
"bytes": 67850,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-msgbox-error-closed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "25e8da8db33e67e7023b10b33dfc3b0fde8c1a416f69b752834e2879e2840e12",
"bytes": 57203,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-msgbox-error-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "823653e5978443f298c60ade95a53e2efcd34822fcf0057f4ced86ddcb40724f",
"bytes": 57403,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-msgbox-info-closed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "036bcd75c028683c240fec3494100dc802914e38465dc2a2b615615488a65d2e",
"bytes": 59345,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-msgbox-info-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e8c67b79d7dbf32df7f1a8cbe725a3d77854549e8d448a65b5d09b09865b4bb5",
"bytes": 59744,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-msgbox-yesno-closed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "095f80354dd993d73e50a8ccd1c0bb33a75d508c666a459011feaa1ed26b4fa4",
"bytes": 65092,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-msgbox-yesno-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bb0a05732941b1cbab81d91c1e3b3edb8b139c5c253c007ba477d24d53ac1d3c",
"bytes": 59734,
"width": 1280,
"height": 1146
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-timer-initial.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d6e308b7403234dad04e106773a1712862d03858d0348a83c054acf360aee910",
"bytes": 42467,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-timer-multiple.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "74ee6563bdabb56539cbd09e765e3f478ec90fc559f522261b84b665202d0cef",
"bytes": 49495,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-timer-running.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "12b1d68aefaf452f806ea155f4f26ff46879a74d048dee748e04c9449ca20145",
"bytes": 48314,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-timer-started.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "3a5190a2eb2fd6c4fd327c291442cb5e5b874440613fb658cb586d0b45c205c5",
"bytes": 45885,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dialogs-timer-stopped.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1846666a73dd41997fd9e39fea78cfe4348a6ddc695b14eb8793905cc78cb72e",
"bytes": 47822,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-02-handlers.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-03-dragenter.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-04-dragleave.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-05-drop.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-06-file-written.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-07-event-fired.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-08-multiple-files.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "dnd-09-with-file.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "037677110c9a53f4d4fac4dcad540e5a54281bae9ccd1f32db89863cce4b9e1e",
"bytes": 35997,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "earlysize-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "14a5404eea7c6d14bcc59d1b54e0e3f981a71391cbcf6571906da0a4a71eda8a",
"bytes": 27335,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "earlysize-02-result.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "14a5404eea7c6d14bcc59d1b54e0e3f981a71391cbcf6571906da0a4a71eda8a",
"bytes": 27335,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eefpsel-01-boot.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c931c622061459244e2acb0e37f53f3048bb992d289cc5024ba5ecb372c1783c",
"bytes": 62646,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-crosshair-00-small.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ef9a510f84b97461009ed758cce2e38ac5ce693f01978a21f6c813f4aeb52969",
"bytes": 57665,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-crosshair-01-full.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "16867e67d2865c493b875b8f66016b34f4c37387c13f2e1c54a7ca786f0c651e",
"bytes": 57918,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-crosshair-02-45.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b1b8124360bf41aa19a40d37a46d2db4d84a7473b9a0f6b4ca6b38689ccddda7",
"bytes": 61814,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-crosshair-03-small-again.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ef9a510f84b97461009ed758cce2e38ac5ce693f01978a21f6c813f4aeb52969",
"bytes": 57665,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-draw-wires-00-before-tool-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c075b77806ba67cb3150244496c204232ce257bf786022f4ab191e7456cf8f15",
"bytes": 55977,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-draw-wires-01-after-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c8ef0c1ebb54a7cba05bde0bbbf402fc8fc58d5a9bda49514ad720035c65e2e7",
"bytes": 57675,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-draw-wires-02-after-drawing.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0c068e0a5b94c203f4aed67e446362ccb33de386c78d461073b67a2e614eb6ff",
"bytes": 58245,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-load-rendered.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "be06d89a4301af681734d06bea4236d8b2b494f5fbf17da5f5b48d8fa3b6f976",
"bytes": 61762,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-loaded-css.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c075b77806ba67cb3150244496c204232ce257bf786022f4ab191e7456cf8f15",
"bytes": 55977,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "eeschema-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c075b77806ba67cb3150244496c204232ce257bf786022f4ab191e7456cf8f15",
"bytes": 55977,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
{
"name": "eeschema-sim-frame.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "f8716d440922de98d596fc02fe636c6c74cac5e39754450a44394dcb205f532f",
"bytes": 87718,
"width": 1280,
"height": 720
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
},
{
"name": "eeschema-sim-plot.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "fbe17f157cf27b5057df83dc7570deff00b4f25ce732cb8d495ae586556cbd3f",
"bytes": 95282,
"width": 1280,
"height": 720
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
},
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
{
"name": "eeschema-url-regex.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "6f5432b5656741d9386d13aafd949bf15735e038b6ad9812dfaaccd15d0f5632",
"bytes": 61974,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "filedialog-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "404f80e779eb2909dad48a942caa7b4d8a9a41d8f3d0338029808b8887fe85dc",
"bytes": 22658,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "filedialog-02-open-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d6aeebf7d70df88b4ac04c4688746d500533d89feb87c524853bb8c5f5c74e61",
"bytes": 37476,
"width": 1280,
"height": 1225
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "filedialog-03-save-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9a3a781de20175409e8d5605405ccdc77eede2a106f333b90b6c0d1e81fe6c74",
"bytes": 38122,
"width": 1280,
"height": 1214
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "filedialog-04-multiple-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "884a0277ec767da898e74e9548da1a42dc51e726da95c674a1e8e6c0a653e4ed",
"bytes": 39225,
"width": 1280,
"height": 1225
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "filedialog-05-all-buttons.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4f3068d49795c8f62cc30a4b975829b6be6df37ec375255f43c706280d76f3dd",
"bytes": 45549,
"width": 1280,
"height": 1225
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "filedlg-folder-nav.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d9703625bd9b4b0b9def909aaaef478ae567082a1384a138800fc9f3a3f75bf3",
"bytes": 37628,
"width": 1280,
"height": 1225
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "fontenum.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c5dcc2b114f0bd003d6cd5e9a6d9eaef8413cc2e2f0fe2a02345a0a731af0c02",
"bytes": 134301,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "frame-runtime-eeschema.html.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c075b77806ba67cb3150244496c204232ce257bf786022f4ab191e7456cf8f15",
"bytes": 55977,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "frame-runtime-footprint_editor.html.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b5400a538dac19a92d8b32f9400ff47a23cd756cab402fc3bdb2d01057962b7c",
"bytes": 94521,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "frame-runtime-pcbnew.html.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "frame-runtime-symbol_editor.html.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "91bd83a6ce37352cbfb6e63e84cdb05e308ee2b51ed81654210127a7400778df",
"bytes": 42706,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gerbview-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "047f32ad58c6b2e6f1d2c31b74c0e9594e984cf0e3b8be223d550aa35d5768df",
"bytes": 38327,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gerbview-02-metrics.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "047f32ad58c6b2e6f1d2c31b74c0e9594e984cf0e3b8be223d550aa35d5768df",
"bytes": 38327,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gerbview-print-00-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4695bb5d42ddb45a95c674654ced552a078e2a5f18151f9bbfa3193c50972f37",
"bytes": 85943,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gerbview-print-01-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "771a22876cc5edc0e1478b47cbf76d0775235e835cca18082fc56fa78489508b",
"bytes": 105906,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gerbview-print-02-reopened.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0802e16507d749f079214ebd3fd12b4eb19e27b1e5853f8b2f2e4a65d303a799",
"bytes": 105905,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "grid-tab-final.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "dbb0f95a668c5ccd98c4b60cc89e0e3e6c0c5fbe1c436083a469fcdc98fa4bc8",
"bytes": 38657,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridedit-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0389f476e026a8b617706d73abf87bf0ce2692284823668f45abd68cc7370d54",
"bytes": 58745,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridedit-02-select-cell.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "71925b1c26df3f87075b241ecb85d37a6cafaf46a2d825ae13ded35183754908",
"bytes": 60189,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridedit-03-edit-cell.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "eef0d57565f49a0a019762cc51a833923446059cc969dc3d4227b880ce143a77",
"bytes": 61936,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridedit-04-add-row.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "63b6e7d9c39ae26dce04e52c2e9e1d7e852e36a117f74addb952da5a68c84904",
"bytes": 61290,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridedit-05-delete-row.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "453abe7e570211a532b677cdca1b8d87fe0830d0c499cba783106978a4721f41",
"bytes": 59252,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridrenderers-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4cafabc637f0f853a87b6123a8bbb80be888361224d2f928b6b47a4887de106e",
"bytes": 39369,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridrenderers-02-color-cells.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4cafabc637f0f853a87b6123a8bbb80be888361224d2f928b6b47a4887de106e",
"bytes": 39369,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridrenderers-03-icon-text.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b4583f75ca74c74d25494f17aa6dc944aeafa775cab82855107ca4dbfb7fe79d",
"bytes": 70015,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridrenderers-04-striped.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8e1263f236dcf036a287749ea002a132d8614d04d1e1fd51d161c6a9d16f98a6",
"bytes": 45026,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "gridrenderers-05-checkbox-toggle.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "944e00a7ad464543de878c42bd3a441df872644d78c90d0e4c6aabfcb38a6904",
"bytes": 45087,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "16a24e90cef87b8cb2c4804170667c3afcac8164e3bd17372a41fc9c88d47896",
"bytes": 49710,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-02-basic-content.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "16a24e90cef87b8cb2c4804170667c3afcac8164e3bd17372a41fc9c88d47896",
"bytes": 49710,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-03-tables.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "14985c6e77bdddb538624c182d88a7a93bba2b60f9991c35abce3b3ceb1ed824",
"bytes": 55093,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-04-long-content.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ee6a3c4e6776845fa44c3b64772976605e6b599813b4b6d29f76070251204799",
"bytes": 73824,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-05-kicad-about.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "05419a02cdb3465f44ef647f74ea5a02282d50ac9de52d570bd1d65e685087f1",
"bytes": 53843,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-06-link-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "cf12518102023a904fe24ec1b922145ff0b7d72755f55052ed8bd61be8ae7a48",
"bytes": 52005,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-07-scrolled.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e5eb4fc45d421fcee4eea0a9f96f98e3d8ba9deb4365ce3abc61ee5541f1fc58",
"bytes": 80340,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-08a-basic.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d7faf13f5a0e3d155cde57eef47770a8b0de2ede91e91f2f63578a408c7ea50f",
"bytes": 52119,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-08b-tables.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "498dd1dcc41e83df06badb9b778d5c3bb9e6427a2cb0a6f30a7b299bbf7b8787",
"bytes": 56926,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "htmlwin-08c-about.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9b457fb87490c2b1268b0cd83139c6a9a718d7e55f49bf756d191295e2b657f7",
"bytes": 57456,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
{
"name": "import-settings-modal-stack.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9d8c0242c934e25ddbc8b07836a972050960efb0626d655c40b02d6f292f1e0a",
"bytes": 148096,
"width": 1280,
"height": 720
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
},
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
{
"name": "infobar-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "442b045e41d3a77de38281f3ea58e89514344384bc0191c87e7bbb4acc01f095",
"bytes": 63306,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "infobar-02-info.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "66fba002067ca9fe3d8516ddac9826a292cdba05776d7be13a48dd945d8463dc",
"bytes": 70159,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "infobar-03-warning.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "862874d6c3794fc996dcca54ee55692b42295d5c58c15f423bf8cad820cd0764",
"bytes": 69899,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "infobar-04-error.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8be3eda782608abb957f55398fdbf432bf3231d92bbf96333b2f736768e9a850",
"bytes": 69037,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "infobar-05-dismiss.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8985031a75c660b899a4661fbdb68cfd8a4ee9076dfeb8ba0f985e4f47f82946",
"bytes": 66274,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "layout-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b6d2e57e0bceba6d43204bd8a9ae2f34d3def0628d30b59cd06c48201cdacbc9",
"bytes": 51899,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "layout-02-splitter.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b6d2e57e0bceba6d43204bd8a9ae2f34d3def0628d30b59cd06c48201cdacbc9",
"bytes": 51899,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "layout-03-sash-dragged.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b1f71756a0a75a43967ab12dc3f9ce4b502916a84c2bb3a797a0411268277613",
"bytes": 54648,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "layout-04-scrolled-left.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b6d2e57e0bceba6d43204bd8a9ae2f34d3def0628d30b59cd06c48201cdacbc9",
"bytes": 51899,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "layout-05-scrolled-right.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b6d2e57e0bceba6d43204bd8a9ae2f34d3def0628d30b59cd06c48201cdacbc9",
"bytes": 51899,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "layout-06-combined.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b1f71756a0a75a43967ab12dc3f9ce4b502916a84c2bb3a797a0411268277613",
"bytes": 54648,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "listctrl-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "28f90168cbeb3f2b1edce35904cdfb208b60244a099d51834b41b48558e12b35",
"bytes": 80345,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "listctrl-02-virtual.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "28f90168cbeb3f2b1edce35904cdfb208b60244a099d51834b41b48558e12b35",
"bytes": 80345,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "listctrl-03-columns.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "28f90168cbeb3f2b1edce35904cdfb208b60244a099d51834b41b48558e12b35",
"bytes": 80345,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "listctrl-04-selected.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "aba59fe63ccd3a908e9a49dbdf43b00424ad707378f4368135cb54d74a02da75",
"bytes": 80788,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "listctrl-05-scrolled.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "dc540b58d96d72ddb97728c4121de566b3e35c433efce74c003b244d8dd6405e",
"bytes": 85449,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "load-pcb-microwave-00-pcbnew-ready.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "load-pcb-microwave-01-dialog-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "84df0436dd6a008fe2d9f700d264d2e268ec3185803b5aefff22ee491af3a9f0",
"bytes": 121801,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "load-pcb-microwave.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c77a82e6fc000bb7d1f875470bae9614ee58ed71d808eb261c6525968b19cc99",
"bytes": 110274,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "load-pcb-pic_programmer-00-pcbnew-ready.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "load-pcb-pic_programmer-01-dialog-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d20cdd3276a509161f0e5a229be73eff8d16ee9a3e4243e6f8a1a469ce87fc5b",
"bytes": 122058,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "load-pcb-pic_programmer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2bb785970e78ac20dafe353a4c628161d6406f517045f9e7674bed75a466dca8",
"bytes": 252106,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "logerror-01-initial.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c208c264da509e122308af765b3955a458846db8c44f93409d7666c04bb12706",
"bytes": 49677,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "logerror-02-before-error.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c208c264da509e122308af765b3955a458846db8c44f93409d7666c04bb12706",
"bytes": 49677,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "logerror-03-single-error-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "df82e13e84a3275c9ff981883acdda7400d9b14703e95ff4a2b7d0303b18bbb5",
"bytes": 69537,
"width": 1280,
"height": 1136
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "logerror-04-multiple-errors-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "89b287c776478cdc697ab5d8a20e344c6d12ebfde1e1d18e9407b35dda0fdfd6",
"bytes": 70165,
"width": 1280,
"height": 1129
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "logerror-05-mixed-levels.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "22980a15cf3b1dfff4833ec9603d2b0a95fb3a686aaab8ef00e937f70272ce0c",
"bytes": 65777,
"width": 1280,
"height": 1129
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "maximize-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "192b41571b6c825396049972e8aedc66c7bbe14f2a4814804d6df5e95c98f20e",
"bytes": 30329,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "maximize-02-fullscreen.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "192b41571b6c825396049972e8aedc66c7bbe14f2a4814804d6df5e95c98f20e",
"bytes": 30329,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "maximize-03-canvas.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "192b41571b6c825396049972e8aedc66c7bbe14f2a4814804d6df5e95c98f20e",
"bytes": 30329,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "menu-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7dd194e7dd380f85f741ef3b6c501dddf661cc1f71e1cbf62aeb6dba8b41f569",
"bytes": 24820,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "menu-02-menubar.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7dd194e7dd380f85f741ef3b6c501dddf661cc1f71e1cbf62aeb6dba8b41f569",
"bytes": 24820,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "menu-03-file-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "da3fce7616932be82861a98405922932ef8883c8c134de02ba2f5209514bf64e",
"bytes": 23929,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "menu-04-edit-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "f72a90ef8ddf7ce94ffb5b5e6efd7730ab4e05667d0a029b1d444d5738a551f7",
"bytes": 25762,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "menu-05-all-menus.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ca1b688681e3c5ddfb776a851c5d7c9272857c452feace02cd4b77b902758a45",
"bytes": 27013,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-chrome.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c6975eb91549be5ccd78915472c5f6fdba4471f5f294bae118101fe1a74352de",
"bytes": 26971,
"width": 412,
"height": 839
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-pan-00-base.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "79acfbe68ced93f72e35959dfbc249d9b050282080941cf0e02d063aa1b5e98c",
"bytes": 26887,
"width": 412,
"height": 839
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-pan-01-panned.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "10fe9d61b4cbacda47c8f8f708ce202ac1b7b72293aee86ae2cec6b668969dc2",
"bytes": 24139,
"width": 412,
"height": 839
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-pan-02-restored.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c6975eb91549be5ccd78915472c5f6fdba4471f5f294bae118101fe1a74352de",
"bytes": 26971,
"width": 412,
"height": 839
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-pinch-00-base.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "f9d0571f1d3b2c63bfac2e3f8fe90f4bd6afee058712035335f07f301f7c9c0d",
"bytes": 26811,
"width": 412,
"height": 839
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-pinch-01-zoomed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9c5d05f907971d9972df5e089d9df9e389148d1c7ba992073f454be412621ff9",
"bytes": 60693,
"width": 412,
"height": 839
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-pinch-02-restored.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "79acfbe68ced93f72e35959dfbc249d9b050282080941cf0e02d063aa1b5e98c",
"bytes": 26887,
"width": 412,
"height": 839
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-toggle-rehidden.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d5c223408d0dd42cd0c2143f8536af619745e1900eab2bf4cfda2c0338d0df57",
"bytes": 32145,
"width": 412,
"height": 839
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "mobile-toggle-shown.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "6394e8cf155fd5c08f538efa4eac9f6fb9e67e563a89d934bfedab5bd72f4f94",
"bytes": 121590,
"width": 412,
"height": 839
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
{
"name": "modal-01-border.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a285bbe8bd3c1dbd20e3e6b013508320c954835c67b04d62362321323724f615",
"bytes": 43477,
"width": 1280,
"height": 922
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "modal-02-before-drag.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7415f2233066c12aefe2cfaa560a281c10c0d3a6f807806095d72ef6402a9e15",
"bytes": 29920,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "modal-03-after-drag.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c63dcfc0d7bdfd8b50ac19b4337337b1df22b133401da42b10dc68c68d30a2a3",
"bytes": 39804,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "modal-04-before-resize.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7415f2233066c12aefe2cfaa560a281c10c0d3a6f807806095d72ef6402a9e15",
"bytes": 29920,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "modal-05-after-resize.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "81f3865dd97c6bad00f06d36d953ed82b0a1eafda9f2ff39bd9007e969a409c2",
"bytes": 25236,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "notebook-01-scrolled.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "6125701515893123955af70ca3908c1f27ca841009bbb1594a364f2a03f2ec33",
"bytes": 29678,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "notebook-02-scrolled-again.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "6125701515893123955af70ca3908c1f27ca841009bbb1594a364f2a03f2ec33",
"bytes": 29678,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "occ-export-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "484b5ad6969f162c0dc344972b957976b8fcf0974f0dbfd09bd049550414acf6",
"bytes": 195991,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "occ-export-done.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e2a72d435d0cc4adfd6a787588961db244d68cd6456d1d3a3ec2d460f2da9599",
"bytes": 188412,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "occ-probe-done.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "fcf0584e219fce4f830acc7c93fd5c76bf63893b0dbd3c20ad1926c79b75e6c7",
"bytes": 6586,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
{
"name": "ownerdrawn-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "659fbe9e53c69209ce50a5463716848cdbfd1b49600574e7be312ec55731a9e3",
"bytes": 40544,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "ownerdrawn-02-layer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "659fbe9e53c69209ce50a5463716848cdbfd1b49600574e7be312ec55731a9e3",
"bytes": 40544,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "ownerdrawn-03-font.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "659fbe9e53c69209ce50a5463716848cdbfd1b49600574e7be312ec55731a9e3",
"bytes": 40544,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "ownerdrawn-04-icon.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "659fbe9e53c69209ce50a5463716848cdbfd1b49600574e7be312ec55731a9e3",
"bytes": 40544,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "ownerdrawn-05-log.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "659fbe9e53c69209ce50a5463716848cdbfd1b49600574e7be312ec55731a9e3",
"bytes": 40544,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pcbnew-context-menu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "214e5854b0176c7d0789557bab76b607590a122a9cca5005dcd4f99a05e54193",
"bytes": 112833,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pcbnew-context-submenu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2a1e9e103070b2a764d36cc80a8c06ac7fd5f27f996cc04ebc1f084ffb8ca13b",
"bytes": 125038,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
{
"name": "pcbnew-dark-mode-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
{
"name": "pcbnew-draw-lines-00-before-tool-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pcbnew-draw-lines-01-after-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a233f32bd68aa497c337f5fc53da02c8920fddca6854383156ad962f65477d52",
"bytes": 107573,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pcbnew-draw-lines-02-after-drawing.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0d45d478dbaeb79e42618a09e9ea3957cba8aa6693037e31e23e2f028c2f5f75",
"bytes": 113101,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pcbnew-loaded-css.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pcbnew-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
{
"name": "pcbnew-move-00-before.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2ce68756b768af4d1e5e8446f839f9f90ccc458dd3db48919e1efdde0cf66d03",
"bytes": 108150,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
{
"name": "pcbnew-move-01-after.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e43612a8ac4c11917b0c8160259d09ec2a5bab640207e70713c3526f3b9e2cec",
"bytes": 111441,
"width": 1280,
"height": 720
fix(tests): enforce screenshot gate; drop shadow dir, flaky retinascale, 12 stale Finalize the screenshot review system after the first Linux re-baseline (that run showed 355/356 stable — the placeholder floor is fine as-is): - ENFORCE the gate: `screenshots:check --fail-on-change` now fails the build on any changed/added/removed vs baselines (it still posts the drift report first, so the failure is actionable; a real render change → re-promote). The generic "CI failed" notice is suppressed for gate failures so we don't double-post on drift. - Exclude retinascale-01-loaded: a fullPage HiDPI test whose captured height + DPR scaling vary run-to-run (~60% inter-run diff) — a flaky test, not render noise. IGNORE_SCREENSHOTS in config.ts; compare/promote/gen-manifest skip it; baseline removed. - Kill the baseline-dir shadowing: drop e2e/baseline-screenshots/ from BASELINE_DIRS + delete its 3 files (grid-tab-final, wxgrid-controls, wxgrid-dedicated-page) that duplicated names in baseline-screenshots/ with different bytes. - Prune 12 stale baselines (renamed/removed specs: wizard-01..04, gerbview-wizard-01..04, zoom-pl_editor-*, popup-03-palette). - Delete the dead compare-screenshots.sh / update-baseline-screenshots.sh. - Regenerate screenshot-manifest.json (355 entries). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-02 11:30:33 +02:00
},
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
{
"name": "pcbnew-sidebar-scrollbar-dragged.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "c14de3d3ff82a6eace168a4d24081fd71d1ede0b76ae4c8b33b057d41ea23e58",
"bytes": 104311,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pcbnew-sidebar-scrollbar.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pickers-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "36311c6e048cb7f367b7478f6bb61d8fa83f86be4c087d4884790e4efc21665c",
"bytes": 44419,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pickers-02-colors.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "36311c6e048cb7f367b7478f6bb61d8fa83f86be4c087d4884790e4efc21665c",
"bytes": 44419,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pickers-03-font.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "36311c6e048cb7f367b7478f6bb61d8fa83f86be4c087d4884790e4efc21665c",
"bytes": 44419,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pickers-04-preview.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "36311c6e048cb7f367b7478f6bb61d8fa83f86be4c087d4884790e4efc21665c",
"bytes": 44419,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "38027e75473ad505f8675fe52f1c7d4e85a42e59b68ba6ea0d93964ed7db6d86",
"bytes": 61645,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-02-no-wizard.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "38027e75473ad505f8675fe52f1c7d4e85a42e59b68ba6ea0d93964ed7db6d86",
"bytes": 61645,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-03-file-menu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d79ddea66737a04a71e136cfaa8c955c70d22fdc7ef5b208901c546c137f7197",
"bytes": 36383,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-04-save-as-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d1a6794ee7d3cd02a1b8fbe4de596f67024824ffe4c12188103da58c34bc1b7",
"bytes": 41910,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-04b-after-enter.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d1a6794ee7d3cd02a1b8fbe4de596f67024824ffe4c12188103da58c34bc1b7",
"bytes": 41910,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-context-menu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "41d1ed13f42b76530f319f5cd76550f6a2e6ee20c40200918c30af97f0de5b7d",
"bytes": 40995,
"width": 1280,
"height": 540
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-load-rendered.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "949caa8b28d50a54bf23e81b30587aea97e8fb01173110169d69ac86eb9e1931",
"bytes": 33734,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-scrollbar-dragged.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9fe5987a042c4c3514512dc9d5490764a9ed9ca51390517d39396da9b92bbdf7",
"bytes": 40589,
"width": 1203,
"height": 460
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "pl_editor-scrollbar.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "f4b0a4b2f84de265b37c67576d2900e1f18a010855a7439698d22dcfd1928948",
"bytes": 40693,
"width": 1203,
"height": 507
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "popup-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "59dd326665bca9781534a591b5a440924a7ce5a73b2e5b2db991840ca49d1044",
"bytes": 39818,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "popup-02-status.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "59dd326665bca9781534a591b5a440924a7ce5a73b2e5b2db991840ca49d1044",
"bytes": 39818,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "popup-03-palette-dismissed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "17057d10f26dc7aa781f416260a1864c160fe6760459ffd703fe2d4489ff089a",
"bytes": 42654,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "popup-03-palette-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d2966206503fed8ee536beb7ac33e8a8119788d30b6db37a5100de1c37c3c929",
"bytes": 40990,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "popup-04-color.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "59dd326665bca9781534a591b5a440924a7ce5a73b2e5b2db991840ca49d1044",
"bytes": 39818,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "popup-05-positioning.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "59dd326665bca9781534a591b5a440924a7ce5a73b2e5b2db991840ca49d1044",
"bytes": 39818,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "popup-06-log.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "59dd326665bca9781534a591b5a440924a7ce5a73b2e5b2db991840ca49d1044",
"bytes": 39818,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "print-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bf1d0278802d2cbb940c330a435fcfa7d6c1ad5ebd30d19d2718e190a45c4dcb",
"bytes": 34760,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "print-02-preview-panel.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bf1d0278802d2cbb940c330a435fcfa7d6c1ad5ebd30d19d2718e190a45c4dcb",
"bytes": 34760,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "print-03-browser-print-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "f606ae66475daf3042c9a947b72660c328b6135d2494088a09fe8706ecd23317",
"bytes": 42574,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "print-04-preview-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ead933f3b167c8fbcfd2aebbcdb9bfa81b96f81f8617142180ff5c2ab6c8c1c1",
"bytes": 38660,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "print-05-page-setup-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "f30638d525e4ca602463d44565d47495417f3803c6794592f190163ce9dc49fd",
"bytes": 37691,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "print-06-print-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b9eb6ee63236cef151812cccae79fd8428f016ac2ba39f276f398f9a56b2eb2c",
"bytes": 36796,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "print-07-callbacks.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ead933f3b167c8fbcfd2aebbcdb9bfa81b96f81f8617142180ff5c2ab6c8c1c1",
"bytes": 38660,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "print-08-all-buttons.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "65af278e5830afc1ce5be1e5f2a3c05bf64e95b9bca41c116512d5bd781d478b",
"bytes": 51497,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "printpreview-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "159a59aa3340ccf15f372176d48d69f73c14be5c6026f34e2c67b8c64e9f6f68",
"bytes": 38659,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "printpreview-02-preview-area.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "159a59aa3340ccf15f372176d48d69f73c14be5c6026f34e2c67b8c64e9f6f68",
"bytes": 38659,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "printpreview-03-preview-frame.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9f7cfd01ed6223bc1753e5b834534a87ee0a98a24d607bb1db5a6ea3cc4e56b8",
"bytes": 61075,
"width": 1280,
"height": 1332
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "printpreview-04-page-setup.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b164f84fdcaa6045bd3a3afe7313d9e80037a59aef1ed65dff949fbb0a7309c5",
"bytes": 60795,
"width": 1280,
"height": 1212
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "printpreview-05-settings.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "159a59aa3340ccf15f372176d48d69f73c14be5c6026f34e2c67b8c64e9f6f68",
"bytes": 38659,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "probe-00-after-wizard.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4ee503ba398bc613a0d243ce0b8ecae01030e8caa348d924acee1f2a125072d5",
"bytes": 106098,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "probe-01-file-menu-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "6d0ec2529f8d7df1071661860bfeeb8c41d296e3c2701a030a96100d283d0c93",
"bytes": 115349,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "probe-02-after-open-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "5b113fbe9f194eda45cea726ef53dbf721b78040f3ddfe4c576ef3c197d6f2ea",
"bytes": 119819,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "probe-03-late.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "5b113fbe9f194eda45cea726ef53dbf721b78040f3ddfe4c576ef3c197d6f2ea",
"bytes": 119819,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
{
"name": "project-sync-00-board-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "548d15e28d12698e01911de97acabeaa111df5abb23a4c6218ebdb4d0285fc23",
"bytes": 95576,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
{
"name": "project-sync-01-update-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8955464503634edf05a335de3d7db9b6b4508d47508af8b371a0363a40e03a97",
"bytes": 130070,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
{
"name": "project-sync-02-resync-first-R777.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e6d94412b4641862d28d75e662bf32ea7f9f3091656581e65527a4517c1788bb",
"bytes": 140967,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
{
"name": "project-sync-03-resync-second-R888.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bd2041a98b1608d2c9e5fe31a973349ee69b049c7c9fc27dc80a78d8c74df81d",
"bytes": 141431,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
{
"name": "propgrid-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "06c56805cf4b5c1c0c47412d8dabdc5cfacd6b4b234d9e94cc2fbb613169ea19",
"bytes": 51560,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "propgrid-02-categories.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "06c56805cf4b5c1c0c47412d8dabdc5cfacd6b4b234d9e94cc2fbb613169ea19",
"bytes": 51560,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "propgrid-03-selected.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "37a31618cbe2ade6e0010412dc2dee376a5a41add0cf1425822510426d949b0a",
"bytes": 53900,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "propgrid-04-manager.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "06c56805cf4b5c1c0c47412d8dabdc5cfacd6b4b234d9e94cc2fbb613169ea19",
"bytes": 51560,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "radiogroups-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "ef78bd25fb753742d37b5dce3e3ab73486aef5e993168a7d973564be964ae0b9",
"bytes": 38359,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "radiogroups-02-selections.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "86704e44570db2289947df6512650d38ef3ccdd8634bb0c0c849bc0e89c999e7",
"bytes": 38377,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "regions.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0660c0d0732fc5a0d323077dafb904358bff290a73f82f47de472bb701a65b52",
"bytes": 63641,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "scrollbar-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9fb3de835c8886b8df5883047bbf0b91c47f7f9763eed01eb4cc255f20832c88",
"bytes": 12039,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "scrollbar-02-dragged.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7425ca16d7277eaa7862b10085a30bfea53cc7759758e17c082ccdea2cb90f9a",
"bytes": 6857,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "searchctrl-01-visible.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bc4ca83b02969d1b2f10c170c6ae2e4e361ac457ca85008233f70c9b31a1e592",
"bytes": 38535,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "searchctrl-02-typing.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "69ea8f9026bfaa4cf220b2cc99b939010e4bc2ef193d4020654087f0a40f27ac",
"bytes": 38692,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "searchctrl-03-after-enter.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "69ea8f9026bfaa4cf220b2cc99b939010e4bc2ef193d4020654087f0a40f27ac",
"bytes": 38692,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "selectheight-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7c4e2dd1d04bae6c077a0ddcd7830f89e37bfd3396ef07e0d64d18aa1eb73877",
"bytes": 15659,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "selectheight-02-result.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7c4e2dd1d04bae6c077a0ddcd7830f89e37bfd3396ef07e0d64d18aa1eb73877",
"bytes": 15659,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "specialized-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1166856ca635c04401e3335e6a30fb463132e71255f6aeb3f40208281952d2fb",
"bytes": 50334,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "specialized-02-treebook-initial.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1166856ca635c04401e3335e6a30fb463132e71255f6aeb3f40208281952d2fb",
"bytes": 50334,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "specialized-03-treebook-subpage.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "71d53fd2c62f719167b9854de0dcb5f815ff2786e4a13da35cbd0ce04f02cc47",
"bytes": 53772,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "specialized-04-treebook-expand.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "fcd5e2e85af9212b748da70600083ea51e832ca08f09624c16681d1f10c15599",
"bytes": 54063,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "specialized-05-bitmapcombo-dropdown.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "fd42e0a04c2777f9c854aa4d0be66c5bd3da05a6dbd9f35cd02bde2e09b2d009",
"bytes": 54818,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "specialized-06-bitmapcombo-select.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "28c2aba2d6724f4f4f0b4bc2ac1f4f3c1ccdca3a7e6d7f88e28a3cf75414c4b2",
"bytes": 54799,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "specialized-07-rearrange-list.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1166856ca635c04401e3335e6a30fb463132e71255f6aeb3f40208281952d2fb",
"bytes": 50334,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "specialized-08-get-order.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "062c704171c91afdb96c6904ecade92563b35e308c2706aed974b7cd3b243eed",
"bytes": 57932,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "spinctrl-01-visible.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bc4ca83b02969d1b2f10c170c6ae2e4e361ac457ca85008233f70c9b31a1e592",
"bytes": 38535,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "spinctrl-02-after-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bc4ca83b02969d1b2f10c170c6ae2e4e361ac457ca85008233f70c9b31a1e592",
"bytes": 38535,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7bfb1550bd13b02bf3b4cc8fdbabe999111ffdda9de1e6b3d46b8beff49fd574",
"bytes": 58176,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-02-python-default.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7bfb1550bd13b02bf3b4cc8fdbabe999111ffdda9de1e6b3d46b8beff49fd574",
"bytes": 58176,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-03-drc-mode.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "284d3097cc6d148a3aba8a590acea1168a47104d52f25f9e6c34ebf2ca369f42",
"bytes": 63365,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-04-plain-mode.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "eccdc3cbf7b2756eec5a91a0906c1d6a6e67310fa26f40fe93b18f6a180e66f4",
"bytes": 58866,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-05-insert-sample.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "0c485809e8d547f6e4b62f176446510d2f88fcec26f2b923b812d1ed06fed56c",
"bytes": 64134,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-06-cleared.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "32058ff074d1013f873cd7a18373db83aa302b6194967dab744394259da257d1",
"bytes": 32610,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-07-line-numbers-toggle.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d23bd16d122f550b08234009accba3adaeeecec78af1d03f259d1cdc1d40b2b7",
"bytes": 57502,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-08-folded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "23161332c6a4606915be386df861f2e37da57b4ec16283776a5f916ce96749ff",
"bytes": 60106,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-09-typed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "dfffbe8f0318ef6734342931db00f6b7e8e47f3852f1a8223638d977475d6c76",
"bytes": 44131,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-10a-python.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7bfb1550bd13b02bf3b4cc8fdbabe999111ffdda9de1e6b3d46b8beff49fd574",
"bytes": 58176,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-10b-drc.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "284d3097cc6d148a3aba8a590acea1168a47104d52f25f9e6c34ebf2ca369f42",
"bytes": 63365,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "stc-10c-python-again.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "21045335a797ea897fb5a11235cc2561efc7144294213c2cc6d8d210290877bd",
"bytes": 63193,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "symbol_editor-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "91bd83a6ce37352cbfb6e63e84cdb05e308ee2b51ed81654210127a7400778df",
"bytes": 42706,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "symbol_editor-02-metrics.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "91bd83a6ce37352cbfb6e63e84cdb05e308ee2b51ed81654210127a7400778df",
"bytes": 42706,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "textdecor.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1299f4d5b9720d7e91365f061cf2e49e8a0d7aba4091ae40de7e6c6197451725",
"bytes": 70599,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "threadpool-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7a43c8836fd5fef111e6dda623c3c78b7419965cf9f5a69cd44b2a31ec374bbe",
"bytes": 31812,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "timer-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "932d43cc87473c7238e3ce6d54411ce3c53524e2204f7ca0368b1e3e418987da",
"bytes": 32107,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "timer-02-started.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9438fd8cc61f22cf4a8b3d3ac36be4eb18ccc9078274008cd351f54ca920278d",
"bytes": 35834,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "timer-03-ticked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "4db6ac851ce90b9ea5aed89c335a675c071978a3a7a1fe74c16f1512cc1e90e3",
"bytes": 35743,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "timer-04-stopped.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "d1aaa5ffa8eb9f6b68a0bab9b918b16124b57573f5c1b6643a1f8e2b434f54a7",
"bytes": 36393,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "timer-05-fast-started.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "53e7586d30edaa41b781eeca592cf618aae86079fe5c48d23c9de2d6af6a4de0",
"bytes": 35583,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "timer-06-fast-running.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b034232114f802580fbce83b90d1f82491fd3c9091c9020f194a8ebddf238490",
"bytes": 35632,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "timer-07-fast-stopped.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9d56d0b2a74cb6ad861ce34f3f330f101299e19bf247a333c28ea8ed121a28aa",
"bytes": 36003,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "timer-08-reset.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2e787910fd221e909fb5b078b63713a6bb1062754d5934dbbec881b6500a3bf8",
"bytes": 34976,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "toolbar-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "edd9f1eecd9d9857bfea90a6cab9f308ab79f885f09fb2d02af4f3413ace82e5",
"bytes": 29039,
"width": 1280,
"height": 721
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "toolbar-02-buttons.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "edd9f1eecd9d9857bfea90a6cab9f308ab79f885f09fb2d02af4f3413ace82e5",
"bytes": 29039,
"width": 1280,
"height": 721
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "toolbar-03-new-clicked.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e54b6dfc44f1d27ae3e889f7d05ae6b49bffbb8b8fea00d0b13f71ec6e01232d",
"bytes": 26639,
"width": 1280,
"height": 721
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "toolbar-04-zoom.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "7aaa32d4a959ee7ce1be79f832706b46e825b1491b4f1ab068f923543151a156",
"bytes": 28654,
"width": 1280,
"height": 721
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "toolbar-05-toggle-on.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9eb1ef157f7e2a9f15a796611b60468601744e796b478a0266f47379f00719be",
"bytes": 26479,
"width": 1280,
"height": 721
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "toolbar-06-toggle-off.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "98a645d91d09ef1cc8fcd950476e043aa5f703a94924a9fa80d82fcbcc65535e",
"bytes": 28138,
"width": 1280,
"height": 721
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "toolbar-07-statusbar.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "edd9f1eecd9d9857bfea90a6cab9f308ab79f885f09fb2d02af4f3413ace82e5",
"bytes": 29039,
"width": 1280,
"height": 721
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "toolbar-08-all-buttons.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1924b802ab54eb08383b7ed05f927c3974d9efcd8654a34c99b6d6c672e7f37f",
"bytes": 35034,
"width": 1280,
"height": 721
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "tree-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8f0462a6d7497ba7a0b6ed66409a290968fe4737381f9a01681f026156e76f5c",
"bytes": 36875,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "tree-02-hierarchy.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8f0462a6d7497ba7a0b6ed66409a290968fe4737381f9a01681f026156e76f5c",
"bytes": 36875,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "tree-03-selected.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "32c87032d668964d9c4fef967328bdc845eedb07b77a81e53ab826cdffdd6234",
"bytes": 39411,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "tree-04-expanded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "46df128a5a60c11e01e0bcd3dd04681c35f27adf0923f159aa074717621aac16",
"bytes": 63260,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "tree-05-collapsed.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "56af815f7d698dd57d38709cd22622eec7ba8384db57af0fe8322ac3f30f8477",
"bytes": 36692,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "tree-06-added.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "6bb014fc49a1a637144f44f95d7f7e012fe9d2acad876b4df794fbf282d741f1",
"bytes": 44180,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "tree-07-deleted.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2ba16f8851c81ef68ba3f48fb21559d59450680fb7d242378514762ee65183f3",
"bytes": 42441,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "validators-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a82d021dc5f6f678643499a732c55c39558c921ded646bcb13e80fd8d5370ba6",
"bytes": 47554,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "validators-02-text.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a82d021dc5f6f678643499a732c55c39558c921ded646bcb13e80fd8d5370ba6",
"bytes": 47554,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "validators-03-integer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a82d021dc5f6f678643499a732c55c39558c921ded646bcb13e80fd8d5370ba6",
"bytes": 47554,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "validators-04-float.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a82d021dc5f6f678643499a732c55c39558c921ded646bcb13e80fd8d5370ba6",
"bytes": 47554,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "validators-05-netname.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a82d021dc5f6f678643499a732c55c39558c921ded646bcb13e80fd8d5370ba6",
"bytes": 47554,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "validators-06-button.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a82d021dc5f6f678643499a732c55c39558c921ded646bcb13e80fd8d5370ba6",
"bytes": 47554,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wasmedge-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d4aa3b94e283e38a389f46c12d3e58b2af1406a1e6b6bfe44c212be14bf4555",
"bytes": 66609,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wasmedge-02-filesystem.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d4aa3b94e283e38a389f46c12d3e58b2af1406a1e6b6bfe44c212be14bf4555",
"bytes": 66609,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wasmedge-03-threading.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d4aa3b94e283e38a389f46c12d3e58b2af1406a1e6b6bfe44c212be14bf4555",
"bytes": 66609,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wasmedge-04-fonts.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d4aa3b94e283e38a389f46c12d3e58b2af1406a1e6b6bfe44c212be14bf4555",
"bytes": 66609,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wasmedge-05-clipboard.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d4aa3b94e283e38a389f46c12d3e58b2af1406a1e6b6bfe44c212be14bf4555",
"bytes": 66609,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wasmedge-06-memory.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d4aa3b94e283e38a389f46c12d3e58b2af1406a1e6b6bfe44c212be14bf4555",
"bytes": 66609,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wasmedge-07-runall.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d4aa3b94e283e38a389f46c12d3e58b2af1406a1e6b6bfe44c212be14bf4555",
"bytes": 66609,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wasmedge-08-log.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8d4aa3b94e283e38a389f46c12d3e58b2af1406a1e6b6bfe44c212be14bf4555",
"bytes": 66609,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-calculator.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9e1b8512ba5e5eb3bbce969891f84fcb40514fd5a052f88b9e162a8109dfc3db",
"bytes": 74004,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-chrome-hidden.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "a75246340fb00484885cfa49e73504432f7a587cd9a2447c42a3f1f289b0fc3e",
"bytes": 82394,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-chrome-restored.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "2cb8d18e415ace9d35938583c1fe2cee8b3143fa22ac006b16900286a58cc94f",
"bytes": 187149,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-eeschema.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1be20a087a6bdf8761b4bee9a7499c6b4d209b77f1732ee0a20080c9f182ff04",
"bytes": 13128,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-footprint_editor.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "858fc1b7f9c879030da3eefd0d167618afee1bebb71f4de4c6d929aeebdbe5da",
"bytes": 73707,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-gerbview.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b347f63f0fdb68851e65551b14791a6621fa8ef93e0a0ecb9975335ad31d17a1",
"bytes": 44869,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
"name": "web-pcbnew.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "339b4fb38e14d81b49527bccebd8e662def076ce03316f33c907f65ba78967c7",
"bytes": 13098,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-pl_editor.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "deccaf9fc3ee77480d930856ac00186644474a6cb8bc18812582f9783d7e4ae2",
"bytes": 12528,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-read-only-viewer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8054e0b4ece2aeb4a0a0ab8db811b48210aee2cd3c894c213286002d51b3324d",
"bytes": 113244,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
{
"name": "web-switch-missing-pcb-created.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b7ca2cd9a46e80337c18717b44a0a5be349d6b163a6ffc78d9284ee2e132bb40",
"bytes": 103962,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
{
"name": "web-switch-pcb-to-sch.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "05122a1eb1881b5826e61a102b45dc1fa1fbb8a6d723e34d8c2e9a73a7676f8e",
"bytes": 77659,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-switch-sch-to-pcb.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "e9eef7e68ff7bacb668f3791e327c15417d90845f2c10729209c458b12a32ea8",
"bytes": 197181,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-symbol_editor.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "b770263b57d9743deb6aaf048f938f1d19deec8e42b8318a366f6264c2437468",
"bytes": 45480,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wizard-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "33ba0f7d0fa58136219b9cf6a58618a9423203cbde525477894b3044bb8034cd",
"bytes": 23601,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wizard-02-launch.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "1f83c1bd8f87b8d63c4078b8e0925a30667294fb04f056353cb16de329caa70b",
"bytes": 46963,
"width": 1280,
"height": 1264
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wizard-03-next-page.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "83a9466d9bf194b9a7396325187115933385a1e3962fc6fe1053c582bf6a04af",
"bytes": 46932,
"width": 1280,
"height": 1264
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wizard-04-back-page.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "83a9466d9bf194b9a7396325187115933385a1e3962fc6fe1053c582bf6a04af",
"bytes": 46932,
"width": 1280,
"height": 1264
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wizard-05-cancel.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "9096123effd77e16e7705ccb54e928cb1fbb690408b8b8c599fcabd4e0ad6f46",
"bytes": 47118,
"width": 1280,
"height": 1264
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wxgrid-01-tab.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "bc4ca83b02969d1b2f10c170c6ae2e4e361ac457ca85008233f70c9b31a1e592",
"bytes": 38535,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wxgrid-02-selection.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "acd5c14ec295f7f4a7a85076c627005bc5634f730a448e4fffe0d326ab7a3cce",
"bytes": 38557,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wxgrid-03-editing.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "acd5c14ec295f7f4a7a85076c627005bc5634f730a448e4fffe0d326ab7a3cce",
"bytes": 38557,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wxgrid-controls.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8f237999b01378a755bb2728002f516fae625d7fd01caf8b75f048a775564b98",
"bytes": 32415,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "wxgrid-dedicated-page.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "8f237999b01378a755bb2728002f516fae625d7fd01caf8b75f048a775564b98",
"bytes": 32415,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "xface-preferences.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "5c7c57d58de052f1c632296e46cb2953679a13918968ab18cf8bc86635366ba2",
"bytes": 149220,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "xml-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "542ae05e3aae73ac18e8c6eac2c4f02ffa27d0d897da0886010bdde8ad1f9d47",
"bytes": 49811,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "xml-02-input.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "542ae05e3aae73ac18e8c6eac2c4f02ffa27d0d897da0886010bdde8ad1f9d47",
"bytes": 49811,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "xml-03-parse.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "542ae05e3aae73ac18e8c6eac2c4f02ffa27d0d897da0886010bdde8ad1f9d47",
"bytes": 49811,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "xml-04-traverse.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "542ae05e3aae73ac18e8c6eac2c4f02ffa27d0d897da0886010bdde8ad1f9d47",
"bytes": 49811,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "xml-05-create.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "542ae05e3aae73ac18e8c6eac2c4f02ffa27d0d897da0886010bdde8ad1f9d47",
"bytes": 49811,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
},
{
"name": "xml-06-results.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "chromium",
"sha256": "542ae05e3aae73ac18e8c6eac2c4f02ffa27d0d897da0886010bdde8ad1f9d47",
"bytes": 49811,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "3d-viewer-00-board-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "7770ead25e3136a0fe9ecc3515c7a57fcbfde42d2ee9bf66965d1a80d0e2fb5f",
"bytes": 304662,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "3d-viewer-models-pic_programmer-render.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "13a3b995487f08ddcc6c1a02296453c31851ed553ed13bafba405690ccd1ce6c",
"bytes": 66250,
"width": 1084,
"height": 623
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "3d-viewer-models-pic_programmer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "699ae900cc1611a9bf1638f346fb2af4aee082d7562feb741ead84b8826da78b",
"bytes": 157000,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "3d-viewer-pic_programmer-render.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "6fb5fc36ad59a4d993984ba27f9cc080bba453c5470b789806f9a92d9fb8abc6",
"bytes": 64238,
"width": 1084,
"height": 623
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "3d-viewer-pic_programmer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "d29f26c34bad9fde6fdbf9b378c42f049a0b32eb196b5e76232e0abb802c8c20",
"bytes": 153293,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "3d-viewer-titlebar.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "cc94b51471f6238b5befbbb65648351579208ab3328823c3a4cad5ca9c8d7911",
"bytes": 153854,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "appearance-00-layers.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "appearance-01-objects.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "743903e05927cefc8247871eba298166cc1df4f2642d0881424f0e4955352f50",
"bytes": 152609,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "appearance-02-nets.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "db8ee4fe7e58e954fbf994ce97d83ac81fec3f4aa38ac742f49f7e0f836477c8",
"bytes": 126338,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "appearance-03-layers-again.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "075de1dfc3786d2b721eff089de55161524cef578254920134278c82e254a345",
"bytes": 152105,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "appearance-10-layers-scrolled.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "1ab1fa82acc129c9d990f7ebc6ec446948966a194966a23fd11b14b9bcc1d454",
"bytes": 151863,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "appearance-11-layers-scrolled-back.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "appearance-20-objects-scrolled.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "ddd293ec09852422d4d4fcd5c52d06315845734222da91953ba677acf94d1496",
"bytes": 154490,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "calculator-before-switch.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "438152d6d9fa39761f32a8df342a7db27175d05124d6fb8141f73a8c1cc3291f",
"bytes": 109471,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "calculator-color-code.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c205d9fbb8d8e9c0174816fca65056924f3b4d33fb6611fc9dda280686981727",
"bytes": 108135,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "calculator-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "438152d6d9fa39761f32a8df342a7db27175d05124d6fb8141f73a8c1cc3291f",
"bytes": 109471,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "coroutine-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "49b10ce00fcbef9a977ed0bdef1fb7a8fef32e9412f55e88a870d538e88d106f",
"bytes": 143828,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "coroutine-02-summary.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "49b10ce00fcbef9a977ed0bdef1fb7a8fef32e9412f55e88a870d538e88d106f",
"bytes": 143828,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "coroutine-nested-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "3c5b72c873bfc2a2ddc350746535317b364e37cfcad0fc5ade498d5df522ea87",
"bytes": 111212,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "coroutine-nested-02-summary.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "3c5b72c873bfc2a2ddc350746535317b364e37cfcad0fc5ade498d5df522ea87",
"bytes": 111212,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eefpsel-01-boot.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "07200fdfd7277f1798bc0ed06cc7615255fbc657d77d1fec1005bd3f3456f46a",
"bytes": 99057,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eefpsel-02-chooser.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "0fb4936a14a7c5958ff671d33f58c36b737721a5ad7fee60eb32751ef72f539c",
"bytes": 117983,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eefpsel-03-selected.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "bd2a388b1715e0cee622687707b1be7b63a862951b5370e7eafd88b42c997afb",
"bytes": 138608,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-crosshair-00-small.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "a22eeffff13a9c0322cd74529facb2f682790efcc71fe4d953d70fb305aedde4",
"bytes": 93739,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-crosshair-01-full.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "880883d92b031d0099579f3bbea04b5790ac87be285096494fd1d41be6ba54d5",
"bytes": 95755,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-crosshair-02-45.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "44176c2be3731babc8ec2b5221e401d065b917a93b47c69dc09ae83443ccb2e2",
"bytes": 99191,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-crosshair-03-small-again.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "a22eeffff13a9c0322cd74529facb2f682790efcc71fe4d953d70fb305aedde4",
"bytes": 93739,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-draw-wires-00-before-tool-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c2721a6679a071285abc4c5b8cd5db66387fb6d5b4384eb62b383d78a7867779",
"bytes": 91180,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-draw-wires-01-after-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "d016eca14a1a3310be7caa652e04092eb1d60fd7aa9fba244777d51491c085ce",
"bytes": 94061,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-draw-wires-02-after-drawing.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "f702c3ebec9c53ea74bc96e17678fe444c882aed7a2222d3d7159452d6db26c9",
"bytes": 94720,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-load-rendered.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c159e2cd0020d29511fa98a424c8a584af8abdd44018e58d7fc73bfaa471f95c",
"bytes": 101522,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-loaded-css.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c2721a6679a071285abc4c5b8cd5db66387fb6d5b4384eb62b383d78a7867779",
"bytes": 91180,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "eeschema-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c2721a6679a071285abc4c5b8cd5db66387fb6d5b4384eb62b383d78a7867779",
"bytes": 91180,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
{
"name": "eeschema-sim-frame.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "8fe0b33efdaf9c3a0a6843c54a4f1508eb68d736a67885d40ffc6a8b5c4b1506",
"bytes": 102937,
"width": 1280,
"height": 720
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
},
{
"name": "eeschema-sim-plot.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "71e9863e4747d14f491fba39269066ca738c678a1ea9bc801064001e7cb85445",
"bytes": 144283,
"width": 1280,
"height": 720
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
},
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
{
"name": "eeschema-url-regex.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "b4ba3159a23a61b31adf087c5c275d519c0fdce891cf41fd385374f3e276cd12",
"bytes": 101034,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "frame-runtime-eeschema.html.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c2721a6679a071285abc4c5b8cd5db66387fb6d5b4384eb62b383d78a7867779",
"bytes": 91180,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "frame-runtime-footprint_editor.html.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "b9ce8aa9d9abbd32728da190ed65a9718201c5021ef67a0f46db93788cafe2ac",
"bytes": 139071,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "frame-runtime-pcbnew.html.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "frame-runtime-symbol_editor.html.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "65e96df0706d45e58752ecd1eade9fe7295d8a19cb4bb7cba0bf73b56621ec85",
"bytes": 71495,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "gerbview-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "70ea0dd9f90d2f77bb37fd8d94d4f8e7ed353e971a5a72a14ca410409dec68be",
"bytes": 65398,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "gerbview-02-metrics.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "70ea0dd9f90d2f77bb37fd8d94d4f8e7ed353e971a5a72a14ca410409dec68be",
"bytes": 65398,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "gerbview-print-00-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "81cfa4f0dc532d873beda50f9ae4cc8d27553975722e7313278d45b8c6d553d4",
"bytes": 120231,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "gerbview-print-01-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "bf113acba1fc65fe6b5e11fd01e9c62b66e386421cb82db3e8262e9fcb7dcdfd",
"bytes": 151967,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "gerbview-print-02-reopened.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "bf113acba1fc65fe6b5e11fd01e9c62b66e386421cb82db3e8262e9fcb7dcdfd",
"bytes": 151967,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
{
"name": "import-settings-modal-stack.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "9fe4d72ad53fd3b2262b4e0d1d2de6640b1c83c57e6509d393c4a13ca0aa19e0",
"bytes": 206463,
"width": 1280,
"height": 720
screenshots: promote CI run 29690403318 baselines Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-19 17:16:58 +02:00
},
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
{
"name": "load-pcb-microwave-00-pcbnew-ready.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "load-pcb-microwave-01-dialog-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "dc0c7e34a3101cd0c54ab250a4f731a82b3b5463b8eda7f9d6c3e49e34f9bb0b",
"bytes": 171687,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "load-pcb-microwave.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "aa69ba3349f5e6d6c608cd34a1ebace79eff118a2ff2a794b36e703a8d4395d0",
"bytes": 153450,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "load-pcb-pic_programmer-00-pcbnew-ready.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "load-pcb-pic_programmer-01-dialog-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "66706697943908b2bc9f89fcdff79b3f60bb6c9ad918d0332e095da2f5fcfae5",
"bytes": 172011,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "load-pcb-pic_programmer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "7770ead25e3136a0fe9ecc3515c7a57fcbfde42d2ee9bf66965d1a80d0e2fb5f",
"bytes": 304662,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "occ-export-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "14cd3d061a423f4e8c0a75793944fb5afdb005048a794b16936cc27397c9ec6e",
"bytes": 261455,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "occ-export-done.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "8a6bef1afcf21d2cfdcc694b0eecf68f831998caedd7542b64a7a147e0e36ad4",
"bytes": 250311,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "occ-probe-done.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "dc062e24106eed983c41d63db637b6bb5be886bb5013959c8ab675a71ff91eeb",
"bytes": 20684,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-context-menu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "03f101c60e72ba7f40fa4a519d2547a0d34cf8813985fc6a7ac6222b1f8da5a1",
"bytes": 163466,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-context-submenu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "86ba7dd4aa1040533eddb906fb689961f25c1ef63213fcd50bbcc227ce3f6891",
"bytes": 183352,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-dark-mode-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-draw-lines-00-before-tool-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-draw-lines-01-after-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "426b98c0c05c80b504df5f441fc39fbe8a58da6b4a20a80ac546417391fe6be2",
"bytes": 154557,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-draw-lines-02-after-drawing.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "1e56a4aaeeb6b0876f1cac6d5473cb87f9757b564c0541686abec2e7110343ad",
"bytes": 162571,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-loaded-css.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-move-00-before.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "7236c3ff14751e8f5b5da5295c693341a9c53141651aad7c34eb62c056a81bdd",
"bytes": 154925,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-move-01-after.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "e7f41335c260d1dd99b706f02c5d20f601241ca669cfbc568232415f0bc40415",
"bytes": 159481,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-sidebar-scrollbar-dragged.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "4f2f476a2c6db6c9ef9919e36329ab3fa5d8a1eb50fabe647c36912919a19ce8",
"bytes": 149279,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pcbnew-sidebar-scrollbar.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "38027e75473ad505f8675fe52f1c7d4e85a42e59b68ba6ea0d93964ed7db6d86",
"bytes": 61645,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-02-no-wizard.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "38027e75473ad505f8675fe52f1c7d4e85a42e59b68ba6ea0d93964ed7db6d86",
"bytes": 61645,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-03-file-menu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "48c4d631fad6e3208fe84ba71dc333c86f4a3814e2fb7cbdadc874c6c15ec31f",
"bytes": 70091,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-04-save-as-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "a38b717669ff3431d45685b2ed10bf1e1e7b82a32ce0fefa8a8759d4fbc543bf",
"bytes": 79049,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-04b-after-enter.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "a38b717669ff3431d45685b2ed10bf1e1e7b82a32ce0fefa8a8759d4fbc543bf",
"bytes": 79049,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-context-menu.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "27b7c26c669660b689153e5ca6af51f01b4a852b9a4ab72822dea359f85c89a3",
"bytes": 75014,
"width": 1280,
"height": 540
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-load-rendered.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "a678f48f3db6917b00fd4527bc05de5c0b5d3ec2def9628b408e921abc0d1c93",
"bytes": 66269,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-scrollbar-dragged.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "d41a2e85f3266e2d8aee67424305e301bdd341d413c7d9c3e80c705adf0651bf",
"bytes": 65874,
"width": 1210,
"height": 460
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "pl_editor-scrollbar.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "e8bd59f74769a46e41aa49c5a6e142907970cc839d1766550959fe8f3e222732",
"bytes": 69769,
"width": 1210,
"height": 523
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "probe-00-after-wizard.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "23df9f84dff056f0778f6b5c0ee1a79f357afa7d4620a1f4f80348793e406bea",
"bytes": 152549,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "probe-01-file-menu-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "a3b5ab2be4356e78c2724441669e55f95019f64f712b61bf56e6aa4094b2815d",
"bytes": 165062,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "probe-02-after-open-click.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "b41631a83cdb97e05d8abb2a4ed3689a854b570e290f379b9afa51806ccfd7ef",
"bytes": 169337,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "probe-03-late.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "b41631a83cdb97e05d8abb2a4ed3689a854b570e290f379b9afa51806ccfd7ef",
"bytes": 169337,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
{
"name": "project-sync-00-board-open.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "81db2eb54e42addd144d68d6e4d1850a763a408a7cc7276317b4ccf405acc090",
"bytes": 141436,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
{
"name": "project-sync-01-update-dialog.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "2e72666dbc4c504cc439cb40a8fe33a6e708c2b9c7fa437cf8aea12d3c62876d",
"bytes": 184510,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
{
"name": "project-sync-02-resync-first-R777.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "f3ac8eaef6fc7b30c5f65332de85e9f731338a91bb4702a7a5a4c1c8466c95b3",
"bytes": 198146,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
{
"name": "project-sync-03-resync-second-R888.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "2f43a8dbdc5ac608396870281c0f2f8a5ab473fbf40f24a59b67cc962b4f72c9",
"bytes": 198596,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
{
"name": "symbol_editor-01-loaded.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "65e96df0706d45e58752ecd1eade9fe7295d8a19cb4bb7cba0bf73b56621ec85",
"bytes": 71495,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "symbol_editor-02-metrics.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "65e96df0706d45e58752ecd1eade9fe7295d8a19cb4bb7cba0bf73b56621ec85",
"bytes": 71495,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-calculator.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "b485f486376ad6981e637b5f7099325663511ce44b69adf9dec5e3a47705c776",
"bytes": 113676,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-chrome-hidden.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "e655f32bd91b8a4d776311cbebdccebdca476e1c7547e8bbe5bfc5521c9a042d",
"bytes": 107437,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-chrome-restored.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "5f9e52c5cde4dde10aea9f1b29254fbdadb8e784cff1addc852d047702e56867",
"bytes": 251747,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-eeschema.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "761a32eb24c87885bac9d3997aaa80a5c18dfe9e084b18d6fb669954eeec0bd5",
"bytes": 117815,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-footprint_editor.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c00bf0e62bc72bab868c13e15b4ce976788f085d4c0598a7f52ea91c8c4f0939",
"bytes": 109074,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-gerbview.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c015870448fde3362b32bf5c723cb90828d9383473fd11dcfe86b0c0a1a3506e",
"bytes": 72286,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-pcbnew.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "08b3cd699ffb1ee702286504e69dfb95077b53ef428114231e9f50022fb34072",
"bytes": 28621,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-pl_editor.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "b58841f992ac4d93dcaae1729bf25a22a8751d0e5e27d0aa8c4b3ef9f0b33aed",
"bytes": 27911,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-read-only-viewer.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "5831b7a8941754109e9d1c0f5216e9189cc6abe64b4718f338b41f56b9c84af2",
"bytes": 145519,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
{
"name": "web-switch-missing-pcb-created.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "c4ef619a19e7bcd46df6bd220521a839007c0a98880373ebf8ea3f3846a3330d",
"bytes": 150635,
"width": 1280,
"height": 720
screenshots: promote CI run 29573746695 baselines 20 updated (17 chromium — SwiftShader→ANGLE render drift; firefox web-eeschema was a stale Loading-splash capture), 10 added (project-sync-00..03 + web-switch-missing-pcb-created, both engines). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Rb9jsqtHsC3tHTaJ45244j
2026-07-17 14:38:02 +02:00
},
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
{
"name": "web-switch-pcb-to-sch.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "300e4a44f78ed6f02dc4716ecb4b85607c84f7d9f3f414ea0a0aa4b167e3cbe5",
"bytes": 119491,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-switch-sch-to-pcb.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "58b0df26f40ed68d1d9a1a48aa5a524eb6bd5233f21fc52d1d167618442fb458",
"bytes": 247284,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "web-symbol_editor.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "252bae8f80ff9042e0e971fab43e2722f7f7ff69598a534dc1266570e99157e2",
"bytes": 74236,
"width": 1280,
"height": 720
e2e/CI: dual-engine suites, per-engine screenshots, SwiftShader retired, prod web suite, CI-coverage gate Squash of experiment/ff-big-modules vs main. Big-module routing removed: native-EH shrank kicad_editor below SpiderMonkey's x86-64 code budget (runs 29355049705/29356152413 green on stock Firefox), so BIG_MODULE_SPECS routing and the baseline-only-JIT crutch are gone — kicad-firefox and kicad-chromium both run the full suite, with the module compiled the way real users' browsers compile it. Per-engine screenshots end to end: stableShot/shotPath write test-results/<engine>/<name>.png; baselines move to baseline-screenshots/{chromium,firefox}/ and the whole tools/screenshots pipeline (compare/promote/manifest/spec-map/changelog/Discord) keys on <engine>/<name>. Previously Firefox and Chromium renders of one spec overwrote each other and Firefox renders were never actually gated. Seeded from CI run 29421380806 (92 new firefox baselines, +24 chromium web-suite shots); manifest generated from the baseline tree. One merged playwright.config.ts (kicad/asyncify/coroutine/perf as projects); ~25 dead npm scripts dropped. The web suite is gated in CI for the first time ever (4 rotted specs fixed, 5 broken lib-bridge specs triaged as fixme in docs/features/web-e2e-rot/); cheap lint step after npm ci; last 26 blind-sleep violations fixed. SwiftShader retired: CI Chromium renders WebGL on ANGLE → Mesa llvmpipe (--use-gl=angle --use-angle=gl --ignore-gpu-blocklist; the blocklist flag is mandatory — llvmpipe is blocklisted and WebGL is silently unavailable without it) in BOTH configs. Under WORKERS=4 congestion SwiftShader transiently failed the first post-board-load draw and the recovery cascade ended in a silent permanent Cairo fallback — that engine flip was the "~1.2% changedRatio both directions" occ-export baseline flake. Validated 160/160 across two 80-repeat rigs; full analysis in docs/features/wx-parity-bugs/occ-export-context-eviction.md. Chromium baselines shift slightly on llvmpipe — promote once from the first green run. Deflakes the new coverage exposed: presence baselines settle before capture; presence fixtures declare current file formats; perf gets its own outputDir so CI evidence survives; occ-export settles the board paint before the export dialog; menu-item waits (waitForRenderedByLabel before clickMenuItem) in 4 specs + the TESTING.md rule. Web suite runs the PROD build, in parallel: webServer becomes backend `start` + the standalone's e2e:preview (build-preview.mjs: link-wasm → stash the public/wasm symlink aside during vite build, build-demo.mjs's move — then vite preview as the persistent server). The wasm middleware serves /wasm/* in preview and emits COOP/COEP/CORP itself (a pthread worker script's own response must carry COEP or Chrome kills it with ERR_BLOCKED_BY_RESPONSE). VITE_* flags bake at build time; VITE_ALLOW_USER_OVERRIDE joins turbo globalEnv. fullyParallel + default workers: 5.2m → 1.4m. Determinism fixes the parallel run exposed: shared-page specs become serial groups; locks.spec grabs alice's exact item via the new kicadCollabTestSelectByUuid hook (cross-tab "first footprint" order is not a ysync invariant); quit specs poll page.url() (quit supersedes its own navigation — NS_BINDING_ABORTED on Firefox). Suite: 51 passed / 12 skipped / 0 failed in 1.6m. CI-coverage gate (lint:ci-coverage): every tests/**/*.spec.ts must be reachable from the npm scripts the workflows invoke — scraped from .github/workflows/, resolved through package.json, coverage asked from playwright --list itself. Rules: uncovered-spec + orphan-project (with a documented LOCAL_ONLY_PROJECTS allowlist). Gating next to lint:determinism; 138 spec files / 13 projects accounted for. Product fixes kept from the investigations (reachable on real GPUs too): wx 7799fd1be5 — paint flags clear before dispatch + Invalidate always propagates; kicad 3dcfea5e45 — SwiftShader pass-boundary flush + per-instance font texture + first-frame GL-error drain (GAL recovery recovers instead of falling back to Cairo) + the user-facing eeschema switch navigates again under __EMSCRIPTEN__ (project-sync's FaceRegistered gate had rerouted it into the hidden sync player; caught by the newly-gated web suite). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_018eUxiPApHgGiu9NFyQfhAq
2026-07-17 12:10:40 +02:00
},
{
"name": "xface-preferences.png",
tests: move screenshot baselines from git to R2 (content-addressed manifest v2) Baseline PNGs (492, 35 MB) leave git: they now live in the private R2 bucket pcbjam-ci-screenshots as immutable sha256/<hex>.png objects, pinned by the committed screenshot-manifest.json (v2: name, engine, sha256, bytes, dims). tests/baseline-screenshots/ becomes a gitignored cache materialized by the new `npm run screenshots:fetch` (r2-sync.ts + aws4fetch r2-store.ts). - promote.ts: requires the RW keypair (shell env or gitignored tests/.env), syncs the cache, uploads new hashes BEFORE rewriting the manifest; the manifest diff is the only git-visible output. --prune only edits the manifest — R2 objects are never deleted, old commits still resolve. - compare.ts: skips the gate (exit 0, no report.json) when the manifest expects baselines but the cache is empty (secretless callers). - changelog.ts: diffs the manifest between revs and fetches bytes from R2; guards against the migration commit (base manifest not v2 → skip). - gen-manifest --check: v2 schema + resurrection guard (fails if baseline PNGs are ever re-committed); credential-free so every caller can gate. - wasm-build.yml: declares optional read-only S3 secrets, caches + fetches baselines before the lint gate; release.yml/deploy-staging.yml pass them. - screenshot-changelog.yml: triggers on the manifest path instead of PNGs. All 492 objects are seeded and hash-verified in the bucket; fetch/compare degrade to a warn-and-skip without credentials. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-08-18 16:01:46 +02:00
"engine": "firefox",
"sha256": "a1667f4697a7dd24e130492a82dccc5647c9c4c468cb21a66eda33f70327923c",
"bytes": 202590,
"width": 1280,
"height": 720
feat(tests): screenshot regression + Discord review, perf-tracked New tooling in tests/tools/screenshots/ (TypeScript via tsx): - compare.ts: one pixelmatch engine (AA-excluded), connected-component "where to look" boxes, old|new+boxes|heatmap triptych, per-engine floors. - promote.ts: churn-free updater — overwrite a baseline only when decoded pixels differ beyond the floor, copying CI bytes verbatim (no re-encode churn); pulls a CI run via `gh run download` or a local --from dir. - post-discord.ts: always-on CI-on-main report (SHA + e2e status + the track-only runtime-perf table), then screenshot triptychs, batched + size-capped + flood-collapsed + 429-aware. - perf-report.ts: perf table with Δ vs the previous main run (via gh). - changelog.ts: no-build git-history baseline differ (Discord trigger B). - noise.ts / gen-manifest.ts: calibration + manifest generation. CI wiring: - wasm-build.yml: post-test step runs the gate + report on the already- produced test-results (no extra build); report-only (continue-on-error), posts only on push to main, inert without DISCORD_WEBHOOK_URL. - ci-ubicloud.yml: secrets: inherit (pass the webhook through). - screenshot-changelog.yml: ~30s no-build changelog on baseline changes. screenshot-manifest.json: canonical 354-name set + best-effort engine tags (313 chromium-swiftshader / 41 firefox-llvmpipe). Normalize scale:'device'->'css' across 18 spec files (no-op at CI DSF=1) so committed baselines are uniformly css-scaled. Design: CI's Linux render is the single source of truth; no pinned container (accept rare env drift -> re-promote); dev commits via promote. Replaces the byte-cmp compare-screenshots.sh + file-size-proxy update-baseline-screenshots.sh (kept for now until the first re-baseline). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-01 18:01:15 +02:00
}
]
}
Reference in a new issue Copy permalink