2025-12-08 12:07:01 +01:00
|
|
|
#!/bin/bash
|
2026-05-29 12:01:44 +02:00
|
|
|
# Build a KiCad app (pcbnew, eeschema, calculator) inside Docker, then run
|
jspi: retire the asyncify pipeline — knob, post-link tail, binaryen hooks
Phase 8 in the parent repo. Deleted: asyncify-scheduler.js, apply-asyncify.sh,
apply-finalize.sh, inject-dyncall-shims.sh, asyncify-imports/removelist.txt,
the wasm-opt/finalize stub pair, scripts/binaryen-hoist-pass/ (the fork stays
a dormant submodule; removal is a follow-up), bench/wasm-opt-bench.sh (README
marked historical), wasm/shims/context_sleep.cpp, and the sched-context
harness app + Makefile targets.
PCBJAM_ASYNC_BACKEND is gone: build-wx-wasm.sh hardcodes the jspi stamp
(still force-cleans pre-migration trees), build-kicad-target.sh gives editors
the JSPI link surface and the CLIs nothing (they pin ASYNCIFY=0), the stub
dance is replaced by an unconditional .real-restore, build-wasm-test.sh lost
its whole post-link loop, docker/build.sh's postprocess is the ENV shim only,
and Makefile.wasm links every app JSPI with the scheduler shim as a tracked
prerequisite. pcbjam_async_policy.h keys on __EMSCRIPTEN__.
jspi-scheduler.js: wxWasmMainLoopPump dropped from the wrap census (the
export died with the D5 detach); inert [TRACE] instrumentation removed.
CI: wasm-build.yml rewritten for the single-cache pipeline (one output cache
keyed on compile inputs; post-processed bytes cached after the shim);
opt_level input removed from both callers. wasm-cache-hash.mjs inputs now
cover patch-env-shim.mjs + jspi-scheduler.js + jspi-exports.txt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 08:39:12 +02:00
|
|
|
# the ENV shim on the host.
|
2026-05-29 06:39:23 +02:00
|
|
|
#
|
|
|
|
|
# Usage:
|
2026-06-11 07:46:48 +02:00
|
|
|
# ./docker/build.sh <app>[,<app>...] [args...]
|
2026-05-29 12:01:44 +02:00
|
|
|
#
|
|
|
|
|
# Apps:
|
2026-07-02 14:48:11 +02:00
|
|
|
# kicad_editor merged PCB + schematic editor image — serves all four editors
|
|
|
|
|
# (PCB / Footprint / Schematic / Symbol) via runtime --frame
|
2026-06-02 13:35:37 +02:00
|
|
|
# calculator PCB calculator
|
|
|
|
|
# pl_editor drawing-sheet editor
|
2026-06-03 14:23:57 +02:00
|
|
|
# gerbview Gerber viewer
|
2026-06-11 07:46:48 +02:00
|
|
|
# all build all of the above
|
2026-07-02 14:48:11 +02:00
|
|
|
# pcbnew standalone PCB engine (debug aid; not deployed — kicad_editor is)
|
|
|
|
|
# eeschema standalone schematic engine (debug aid; not deployed)
|
2026-06-11 07:46:48 +02:00
|
|
|
#
|
|
|
|
|
# A comma-separated list builds just those apps in order (e.g.
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# "calculator,pl_editor" — used to exercise the multi-app path cheaply).
|
2026-05-29 12:01:44 +02:00
|
|
|
#
|
|
|
|
|
# Any extra args are forwarded to scripts/kicad/build-<app>.sh (e.g. -j 8,
|
|
|
|
|
# --full, --release, --diag=gal).
|
2025-12-13 22:13:31 +01:00
|
|
|
#
|
|
|
|
|
# The build is split into two phases:
|
jspi: retire the asyncify pipeline — knob, post-link tail, binaryen hooks
Phase 8 in the parent repo. Deleted: asyncify-scheduler.js, apply-asyncify.sh,
apply-finalize.sh, inject-dyncall-shims.sh, asyncify-imports/removelist.txt,
the wasm-opt/finalize stub pair, scripts/binaryen-hoist-pass/ (the fork stays
a dormant submodule; removal is a follow-up), bench/wasm-opt-bench.sh (README
marked historical), wasm/shims/context_sleep.cpp, and the sched-context
harness app + Makefile targets.
PCBJAM_ASYNC_BACKEND is gone: build-wx-wasm.sh hardcodes the jspi stamp
(still force-cleans pre-migration trees), build-kicad-target.sh gives editors
the JSPI link surface and the CLIs nothing (they pin ASYNCIFY=0), the stub
dance is replaced by an unconditional .real-restore, build-wasm-test.sh lost
its whole post-link loop, docker/build.sh's postprocess is the ENV shim only,
and Makefile.wasm links every app JSPI with the scheduler shim as a tracked
prerequisite. pcbjam_async_policy.h keys on __EMSCRIPTEN__.
jspi-scheduler.js: wxWasmMainLoopPump dropped from the wrap census (the
export died with the D5 detach); inert [TRACE] instrumentation removed.
CI: wasm-build.yml rewritten for the single-cache pipeline (one output cache
keyed on compile inputs; post-processed bytes cached after the shim);
opt_level input removed from both callers. wasm-cache-hash.mjs inputs now
cover patch-env-shim.mjs + jspi-scheduler.js + jspi-exports.txt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 08:39:12 +02:00
|
|
|
# 1. Docker: Compile KiCad to WASM (fully finalized; JSPI links in-container)
|
|
|
|
|
# 2. Host: ENV merge shim on the glue (patch-env-shim.mjs)
|
2025-12-13 22:13:31 +01:00
|
|
|
|
2026-06-01 17:31:03 +02:00
|
|
|
# Auto-launch the live progress dashboard in this terminal (handled by logging.sh,
|
|
|
|
|
# which owns the TTY before it re-execs us with output redirected). Set KICAD_NO_MONITOR=1
|
|
|
|
|
# to disable. MUST be set before sourcing logging.sh — that's where the dashboard is
|
|
|
|
|
# launched, in the pre-re-exec process.
|
|
|
|
|
export KICAD_MONITOR=1
|
|
|
|
|
|
2026-05-29 12:01:44 +02:00
|
|
|
# Redirect all output to a log file (re-execs script with redirection).
|
|
|
|
|
# MUST be sourced before arg parsing — the re-exec relies on the original
|
|
|
|
|
# "$@", so any shifts before this point would strip args from the re-exec.
|
2026-05-29 06:39:23 +02:00
|
|
|
source "$(dirname "$0")/../scripts/common/logging.sh"
|
|
|
|
|
|
2026-06-01 17:31:03 +02:00
|
|
|
# Build-progress markers (parsed by scripts/build-monitor.sh).
|
|
|
|
|
source "$(dirname "$0")/../scripts/common/stages.sh"
|
|
|
|
|
|
2026-06-29 19:50:18 +02:00
|
|
|
# Pinned toolchain version (single source of truth). Exported so the compose build.args can pass it
|
|
|
|
|
# into the Docker image's emsdk install — bumping the toolchain is then a one-line edit in versions.sh.
|
|
|
|
|
source "$(dirname "$0")/../scripts/common/versions.sh"
|
|
|
|
|
export EMSCRIPTEN_VERSION
|
|
|
|
|
|
2025-12-08 12:07:01 +01:00
|
|
|
set -e
|
|
|
|
|
|
2026-07-03 14:42:38 +02:00
|
|
|
# Stop the builder container when the build ends (any exit path). A lingering
|
|
|
|
|
# idle builder keeps the Docker Desktop VM ballooned — each project's builder
|
|
|
|
|
# is capped at ${KICAD_DOCKER_MEM:-32G}, and with per-worktree compose projects
|
|
|
|
|
# two forgotten containers once read as ~56 GB of host RAM (page cache +
|
|
|
|
|
# high-water ballooning). The container is pure scaffolding: all caches live in
|
|
|
|
|
# the named volumes and `up -d` restarts it in seconds. Set
|
|
|
|
|
# KICAD_KEEP_CONTAINER=1 to keep it running for interactive exec/debugging.
|
|
|
|
|
_COMPOSE_FILE="$(cd "$(dirname "$0")" && pwd)/docker-compose.yml"
|
|
|
|
|
stop_builder() {
|
|
|
|
|
if [[ "${KICAD_KEEP_CONTAINER:-0}" != "1" ]]; then
|
|
|
|
|
docker compose -f "${_COMPOSE_FILE}" stop >/dev/null 2>&1 || true
|
|
|
|
|
fi
|
|
|
|
|
}
|
|
|
|
|
|
2026-06-01 17:31:03 +02:00
|
|
|
# Emit a completion/failure marker no matter how the build ends, so the monitor
|
|
|
|
|
# can stop on a clean "done" or show an aborted state instead of hanging.
|
2026-07-03 14:42:38 +02:00
|
|
|
trap '_rc=$?; if [ $_rc -eq 0 ]; then kw_done; else kw_fail $_rc; fi; stop_builder' EXIT
|
2026-06-01 17:31:03 +02:00
|
|
|
# On Ctrl-C, mark the build aborted so the final dashboard frame shows failed
|
|
|
|
|
# (not a stale "running" state). The EXIT trap above also fires; the monitor reads
|
|
|
|
|
# the last marker, so the duplicate is harmless.
|
|
|
|
|
trap 'kw_fail 130; exit 130' INT TERM
|
|
|
|
|
|
2025-12-08 12:07:01 +01:00
|
|
|
cd "$(dirname "$0")/.."
|
|
|
|
|
|
2026-07-19 15:59:21 +02:00
|
|
|
VALID_APPS="kicad_editor | pcbnew | eeschema | calculator | pl_editor | gerbview | kicad_tools | occ_service | ngspice_service | all"
|
2026-05-29 12:01:44 +02:00
|
|
|
|
|
|
|
|
usage() {
|
2026-06-11 07:46:48 +02:00
|
|
|
echo "Usage: ./docker/build.sh <app>[,<app>...] [args...]" >&2
|
2026-05-29 12:01:44 +02:00
|
|
|
echo " <app>: ${VALID_APPS}" >&2
|
|
|
|
|
echo " args: forwarded to scripts/kicad/build-<app>.sh (e.g. -j 8, --release)" >&2
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
# First positional arg must be the app name. No default — picking one would
|
|
|
|
|
# silently build the wrong thing for someone who forgot the argument.
|
|
|
|
|
if [[ "${1:-}" == "-h" ]] || [[ "${1:-}" == "--help" ]]; then
|
|
|
|
|
usage
|
|
|
|
|
exit 0
|
|
|
|
|
fi
|
|
|
|
|
if [[ $# -lt 1 ]] || [[ "$1" == -* ]]; then
|
|
|
|
|
echo "Error: missing <app> argument" >&2
|
|
|
|
|
usage
|
|
|
|
|
exit 1
|
2026-05-29 06:39:23 +02:00
|
|
|
fi
|
2026-05-29 12:01:44 +02:00
|
|
|
APP_NAME="$1"
|
|
|
|
|
shift
|
2026-05-29 06:39:23 +02:00
|
|
|
|
2026-06-11 07:46:48 +02:00
|
|
|
# Expand the app argument into APPS[]: "all", a single app, or a comma list.
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# kicad_editor first in "all" — the merged image is the deployed bundle and the
|
|
|
|
|
# longest compile, so it starts first and surfaces failures earliest.
|
|
|
|
|
# pcbnew/eeschema stay buildable as standalone debug aids but are not part of
|
|
|
|
|
# "all" (not deployed). kicad_tools joined "all" for the runner-image CI
|
|
|
|
|
# (tasks-runner 0001 R2).
|
2026-06-11 07:46:48 +02:00
|
|
|
if [[ "$APP_NAME" == "all" ]]; then
|
2026-07-19 15:59:21 +02:00
|
|
|
APPS=(kicad_editor occ_service ngspice_service calculator pl_editor gerbview kicad_tools)
|
2026-06-11 07:46:48 +02:00
|
|
|
else
|
|
|
|
|
IFS=',' read -r -a APPS <<< "$APP_NAME"
|
|
|
|
|
for app in "${APPS[@]}"; do
|
|
|
|
|
case "$app" in
|
2026-07-19 15:59:21 +02:00
|
|
|
kicad_editor|pcbnew|eeschema|calculator|pl_editor|gerbview|kicad_tools|occ_service|ngspice_service) ;;
|
2026-06-11 07:46:48 +02:00
|
|
|
*)
|
|
|
|
|
echo "Error: unknown app '$app' (expected: ${VALID_APPS})" >&2
|
|
|
|
|
usage
|
|
|
|
|
exit 1
|
|
|
|
|
;;
|
|
|
|
|
esac
|
|
|
|
|
done
|
|
|
|
|
fi
|
2026-05-29 06:39:23 +02:00
|
|
|
|
2026-06-01 16:30:29 +02:00
|
|
|
# Use branch name as Docker Compose project name for isolated containers/volumes.
|
|
|
|
|
# Honor a pre-set COMPOSE_PROJECT_NAME so a build can target an existing volume
|
|
|
|
|
# (e.g. reuse another branch's already-provisioned deps).
|
2026-01-03 07:31:46 +01:00
|
|
|
BRANCH_NAME=$(git rev-parse --abbrev-ref HEAD | tr '/' '-' | tr '[:upper:]' '[:lower:]')
|
2026-06-01 16:30:29 +02:00
|
|
|
export COMPOSE_PROJECT_NAME="${COMPOSE_PROJECT_NAME:-kicad-wasm-${BRANCH_NAME}}"
|
2026-01-03 07:31:46 +01:00
|
|
|
echo "Using Docker project: ${COMPOSE_PROJECT_NAME}"
|
2026-05-29 06:39:23 +02:00
|
|
|
echo "Building app: ${APP_NAME}"
|
2026-01-03 07:31:46 +01:00
|
|
|
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# Build phase (cache split). The container compile emits the finalized wasm;
|
|
|
|
|
# the host tail is only the ENV merge shim on the glue. Splitting compile and
|
|
|
|
|
# postprocess lets CI cache the expensive compile once and re-run just the
|
|
|
|
|
# host tail — see .github/workflows/.
|
2026-06-22 14:07:50 +02:00
|
|
|
# (default) both — compile in-container, then host post-process.
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# --compile-only — only the in-container compile → wasm+glue in output/.
|
jspi: retire the asyncify pipeline — knob, post-link tail, binaryen hooks
Phase 8 in the parent repo. Deleted: asyncify-scheduler.js, apply-asyncify.sh,
apply-finalize.sh, inject-dyncall-shims.sh, asyncify-imports/removelist.txt,
the wasm-opt/finalize stub pair, scripts/binaryen-hoist-pass/ (the fork stays
a dormant submodule; removal is a follow-up), bench/wasm-opt-bench.sh (README
marked historical), wasm/shims/context_sleep.cpp, and the sched-context
harness app + Makefile targets.
PCBJAM_ASYNC_BACKEND is gone: build-wx-wasm.sh hardcodes the jspi stamp
(still force-cleans pre-migration trees), build-kicad-target.sh gives editors
the JSPI link surface and the CLIs nothing (they pin ASYNCIFY=0), the stub
dance is replaced by an unconditional .real-restore, build-wasm-test.sh lost
its whole post-link loop, docker/build.sh's postprocess is the ENV shim only,
and Makefile.wasm links every app JSPI with the scheduler shim as a tracked
prerequisite. pcbjam_async_policy.h keys on __EMSCRIPTEN__.
jspi-scheduler.js: wxWasmMainLoopPump dropped from the wrap census (the
export died with the D5 detach); inert [TRACE] instrumentation removed.
CI: wasm-build.yml rewritten for the single-cache pipeline (one output cache
keyed on compile inputs; post-processed bytes cached after the shim);
opt_level input removed from both callers. wasm-cache-hash.mjs inputs now
cover patch-env-shim.mjs + jspi-scheduler.js + jspi-exports.txt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 08:39:12 +02:00
|
|
|
# --postprocess-only — only the host post-process (ENV merge shim) on the
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# existing output/ glue; NO container needed.
|
2026-06-22 14:07:50 +02:00
|
|
|
# Extracted here so they are NOT forwarded to the inner build-<app>.sh scripts.
|
|
|
|
|
PHASE="both"
|
|
|
|
|
_FILTERED=()
|
|
|
|
|
for _arg in "$@"; do
|
|
|
|
|
case "$_arg" in
|
|
|
|
|
--compile-only) PHASE="compile" ;;
|
|
|
|
|
--postprocess-only) PHASE="postprocess" ;;
|
|
|
|
|
*) _FILTERED+=("$_arg") ;;
|
|
|
|
|
esac
|
|
|
|
|
done
|
|
|
|
|
set -- "${_FILTERED[@]+"${_FILTERED[@]}"}"
|
|
|
|
|
|
2025-12-15 11:47:50 +01:00
|
|
|
# Add -j 10 by default if no -j flag is given
|
|
|
|
|
ARGS=("$@")
|
|
|
|
|
if [[ ! " ${ARGS[*]} " =~ " -j " ]]; then
|
|
|
|
|
ARGS+=("-j" "10")
|
|
|
|
|
fi
|
|
|
|
|
|
2026-06-22 14:07:50 +02:00
|
|
|
# Container compile + source sync — skipped entirely for --postprocess-only,
|
|
|
|
|
# which is pure host work on the already-built base wasm in output/.
|
|
|
|
|
if [[ "$PHASE" != "postprocess" ]]; then
|
|
|
|
|
|
2026-06-29 19:50:18 +02:00
|
|
|
# Start container if not running. --build so the image is rebuilt when the pinned EMSCRIPTEN_VERSION
|
|
|
|
|
# (build-arg from versions.sh) changes; Docker layer-caches it to a near no-op when unchanged.
|
|
|
|
|
docker compose -f docker/docker-compose.yml up -d --build
|
2025-12-08 12:07:01 +01:00
|
|
|
|
2026-01-05 10:20:08 +01:00
|
|
|
# Sync source code to container volume (fixes macOS Docker VirtioFS issues)
|
|
|
|
|
# Use --checksum to only transfer files with different CONTENT, not timestamps.
|
|
|
|
|
# This avoids the timestamp mismatch cycle that caused full rebuilds every time.
|
|
|
|
|
# Transferred files get current container time, so make detects them correctly.
|
2026-06-01 17:31:03 +02:00
|
|
|
kw_stage container-sync
|
jspi: migration phases 0-7 — build knob, scheduler shim, test successor suite
Toolchain: emsdk 6.0.6 (versions.sh; cache-hash keys on it). Build knob
PCBJAM_ASYNC_BACKEND=jspi|asyncify: build-kicad-target.sh links editors with
-sJSPI + -sJSPI_EXPORTS=@scripts/common/jspi-exports.txt + --pre-js
jspi-scheduler.js (no DYNCALLS, no post-link asyncify pipeline); wx build
stamps the backend and forces clean on flip or unknown provenance;
docker/build.sh passes the knob, seeds the emscripten ports cache from the
volume every launch, jspi postprocess = patch-env-shim only.
scripts/common/shims/jspi-scheduler.js: the JSPI successor scheduler —
token-wait registry, resume turnstile (one armed resume between engine
re-entries, SP swaps only at microtask boundaries), green-region spill
stacks (16-aligned tops), S1 embind mutator FIFO lane + parker wraps, S6
shutdown, libctx integration hooks (suspend/end/quarantine + g_current
arm/clear), SuspendError attributor, lost-wake + stuck-window watchdogs,
__wxWaitDump observability.
Embind: PARKER registrations get emscripten::async() under PCBJAM_JSPI
(wasm/bindings/pcbjam_async_policy.h). nanosleep yields route via the shim.
Tests: tests/asyncify -> tests/jspi successor suite (jspi-stack red/green
shadow-stack battery, jspi-coroutine MiniCoro harness, suspend-races
semantic scenarios + __wxWaitDump books coherence); projects jspi-firefox/
jspi-chrome (asyncify-webkit retired — no JSPI in WebKit); unconditional
Firefox JSPI pref; guard-beacons -> wait-beacons (+wxScheduler/libctxJspi
families); Makefile.wasm links test apps against JSPI with the shim as a
tracked link prerequisite.
Web: WasmTool setRo await + __wxWaitDump forensics, open-flow contained
promise, scheduler-shim.test.ts retargeted (8 green).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 07:06:24 +02:00
|
|
|
# Re-seed the emscripten ports cache from the persistent volume: /emsdk lives
|
|
|
|
|
# in the container layer, so any image-context edit recreates the container
|
|
|
|
|
# and wipes the cache — and the in-container github fetch is flaky. Seeds are
|
|
|
|
|
# staged once into the kicad-build-cache volume (emcache/ports); this copy
|
|
|
|
|
# survives every recreation. No-op when the staging dir is absent.
|
|
|
|
|
docker compose -f docker/docker-compose.yml exec kicad-wasm-builder bash -c \
|
|
|
|
|
'if [ -d /workspace/build-wasm/emcache/ports ]; then \
|
|
|
|
|
mkdir -p /emsdk/upstream/emscripten/cache/ports && \
|
|
|
|
|
cp -a /workspace/build-wasm/emcache/ports/. /emsdk/upstream/emscripten/cache/ports/ && \
|
|
|
|
|
echo "seeded emscripten ports cache from volume"; fi' || true
|
2025-12-27 01:07:44 +01:00
|
|
|
echo "Syncing source code to container..."
|
2026-05-28 17:21:26 +02:00
|
|
|
# rsync into the macOS-backed volume intermittently hits transient VirtioFS glitches:
|
|
|
|
|
# temp-file rename failures (exit 23) or vanished-source files (exit 24, harmless).
|
|
|
|
|
# --inplace avoids the temp-file+rename pattern that triggers exit 23; retry up to 3x
|
|
|
|
|
# for any residual flakiness (--checksum makes each retry skip already-synced files).
|
|
|
|
|
sync_rc=0
|
|
|
|
|
for sync_attempt in 1 2 3; do
|
|
|
|
|
if docker compose -f docker/docker-compose.yml exec kicad-wasm-builder \
|
|
|
|
|
rsync -r --delete --checksum --inplace \
|
|
|
|
|
--exclude="build-wasm" \
|
|
|
|
|
--exclude="output" \
|
|
|
|
|
--exclude=".git" \
|
|
|
|
|
--exclude="logs" \
|
|
|
|
|
--exclude=".idea" \
|
|
|
|
|
--exclude="node_modules" \
|
|
|
|
|
--exclude="tools/emsdk" \
|
|
|
|
|
/workspace-host/ /workspace/
|
|
|
|
|
then
|
|
|
|
|
sync_rc=0
|
|
|
|
|
else
|
|
|
|
|
sync_rc=$?
|
|
|
|
|
fi
|
|
|
|
|
{ [ $sync_rc -eq 0 ] || [ $sync_rc -eq 24 ]; } && break
|
|
|
|
|
echo "rsync attempt ${sync_attempt} failed (exit ${sync_rc}); retrying in 2s..."
|
|
|
|
|
sleep 2
|
|
|
|
|
done
|
|
|
|
|
if [ $sync_rc -ne 0 ] && [ $sync_rc -ne 24 ]; then
|
|
|
|
|
echo "ERROR: source sync failed after retries (exit ${sync_rc})"; exit 1
|
|
|
|
|
fi
|
2025-12-27 01:07:44 +01:00
|
|
|
|
2026-06-22 14:07:50 +02:00
|
|
|
fi # end: container compile + sync guard ("$PHASE" != postprocess)
|
|
|
|
|
|
2026-05-29 12:01:44 +02:00
|
|
|
# Map an app name to its inner CMake build subdirectory. Most apps share their
|
|
|
|
|
# subdir name with the app name; pcb_calculator emits OUTPUT_NAME=calculator
|
|
|
|
|
# but lives under the pcb_calculator/ subtree.
|
|
|
|
|
kicad_subdir_for() {
|
|
|
|
|
case "$1" in
|
2026-06-14 06:22:40 +02:00
|
|
|
calculator) echo "pcb_calculator" ;;
|
|
|
|
|
pl_editor) echo "pagelayout_editor" ;;
|
|
|
|
|
*) echo "$1" ;;
|
2026-05-29 12:01:44 +02:00
|
|
|
esac
|
|
|
|
|
}
|
|
|
|
|
|
2026-06-11 07:46:48 +02:00
|
|
|
# Phase 1 of one app: compile in the container and copy the output to ./output.
|
2026-06-01 17:31:03 +02:00
|
|
|
# Args: <app> [index] [total] — index/total drive the monitor's app counter.
|
2026-06-11 07:46:48 +02:00
|
|
|
compile_app() {
|
2026-05-29 06:39:23 +02:00
|
|
|
local app="$1"
|
2026-06-01 17:31:03 +02:00
|
|
|
local index="${2:-1}"
|
|
|
|
|
local total="${3:-1}"
|
2026-05-29 12:01:44 +02:00
|
|
|
local subdir
|
|
|
|
|
subdir=$(kicad_subdir_for "$app")
|
2026-06-01 17:31:03 +02:00
|
|
|
kw_app "$app" "$index" "$total"
|
2026-05-29 06:39:23 +02:00
|
|
|
echo ""
|
2026-06-01 17:31:03 +02:00
|
|
|
echo "=== Building ${app} (${index}/${total}) ==="
|
2026-05-29 06:39:23 +02:00
|
|
|
|
2026-06-12 10:18:16 +02:00
|
|
|
local out_dir="output"
|
2026-06-11 00:19:53 +02:00
|
|
|
local kicad_build="kicad-${app}"
|
|
|
|
|
|
2026-05-29 12:01:44 +02:00
|
|
|
# Run build inside the container.
|
2026-05-29 06:39:23 +02:00
|
|
|
# -e EMSDK=/emsdk: `docker compose exec` bypasses the entrypoint that sources
|
2026-05-29 12:01:44 +02:00
|
|
|
# emsdk_env.sh, so the build shell would lack emcc/embuilder on PATH. Setting
|
|
|
|
|
# EMSDK lets scripts/common/env.sh source /emsdk/emsdk_env.sh and activate the toolchain.
|
2026-07-10 13:13:51 +02:00
|
|
|
# BUILD_3D_VIEWER passes through EMPTY when the host didn't set it, so
|
|
|
|
|
# build-kicad-target.sh can apply per-app defaults (ON for editors, OFF
|
2026-07-10 17:09:35 +02:00
|
|
|
# for headless CLIs like kicad_tools — the gl1 shim needs glm).
|
2026-06-12 10:18:16 +02:00
|
|
|
docker compose -f docker/docker-compose.yml exec -e EMSDK=/emsdk \
|
2026-07-10 13:13:51 +02:00
|
|
|
-e BUILD_3D_VIEWER="${BUILD_3D_VIEWER:-}" \
|
2026-06-11 00:19:53 +02:00
|
|
|
kicad-wasm-builder \
|
2026-05-29 06:39:23 +02:00
|
|
|
"/workspace/scripts/kicad/build-${app}.sh" "${ARGS[@]}"
|
|
|
|
|
|
|
|
|
|
# Copy output to host-accessible directory.
|
|
|
|
|
# ${app}.wasm.debug.wasm contains DWARF debug info (when built with -gseparate-dwarf).
|
2026-06-01 17:31:03 +02:00
|
|
|
kw_stage copy-output
|
2026-06-11 00:19:53 +02:00
|
|
|
echo "Copying ${app} build output to ./${out_dir}/..."
|
2026-05-29 06:39:23 +02:00
|
|
|
docker compose -f docker/docker-compose.yml exec kicad-wasm-builder \
|
2026-06-11 00:19:53 +02:00
|
|
|
bash -c "mkdir -p /workspace/${out_dir} && \
|
|
|
|
|
cp /workspace/build-wasm/${kicad_build}/${subdir}/${app}.{js,wasm,wasm.debug.wasm,wasm.map,worker.js} /workspace/${out_dir}/ 2>/dev/null || \
|
|
|
|
|
cp /workspace/build-wasm/${kicad_build}/${subdir}/${app}.{js,wasm} /workspace/${out_dir}/; \
|
|
|
|
|
cp /workspace/build-wasm/${kicad_build}/resources/images.tar.gz /workspace/${out_dir}/ 2>/dev/null || true; \
|
2026-06-12 10:18:16 +02:00
|
|
|
cp /workspace/wxwidgets/build/wasm/wx.js /workspace/${out_dir}/ 2>/dev/null || true; \
|
|
|
|
|
cp /workspace/wxwidgets/build/wasm/wx-dom.js /workspace/${out_dir}/ 2>/dev/null || true"
|
2026-05-29 06:39:23 +02:00
|
|
|
|
2026-06-11 07:46:48 +02:00
|
|
|
# The container runs as root, so files in the bind-mounted ./output land
|
|
|
|
|
# root-owned on the host. macOS Docker Desktop remaps ownership to the host
|
jspi: retire the asyncify pipeline — knob, post-link tail, binaryen hooks
Phase 8 in the parent repo. Deleted: asyncify-scheduler.js, apply-asyncify.sh,
apply-finalize.sh, inject-dyncall-shims.sh, asyncify-imports/removelist.txt,
the wasm-opt/finalize stub pair, scripts/binaryen-hoist-pass/ (the fork stays
a dormant submodule; removal is a follow-up), bench/wasm-opt-bench.sh (README
marked historical), wasm/shims/context_sleep.cpp, and the sched-context
harness app + Makefile targets.
PCBJAM_ASYNC_BACKEND is gone: build-wx-wasm.sh hardcodes the jspi stamp
(still force-cleans pre-migration trees), build-kicad-target.sh gives editors
the JSPI link surface and the CLIs nothing (they pin ASYNCIFY=0), the stub
dance is replaced by an unconditional .real-restore, build-wasm-test.sh lost
its whole post-link loop, docker/build.sh's postprocess is the ENV shim only,
and Makefile.wasm links every app JSPI with the scheduler shim as a tracked
prerequisite. pcbjam_async_policy.h keys on __EMSCRIPTEN__.
jspi-scheduler.js: wxWasmMainLoopPump dropped from the wrap census (the
export died with the D5 detach); inert [TRACE] instrumentation removed.
CI: wasm-build.yml rewritten for the single-cache pipeline (one output cache
keyed on compile inputs; post-processed bytes cached after the shim);
opt_level input removed from both callers. wasm-cache-hash.mjs inputs now
cover patch-env-shim.mjs + jspi-scheduler.js + jspi-exports.txt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 08:39:12 +02:00
|
|
|
# user, but on a Linux CI runner the host-side ENV-shim step can't write
|
|
|
|
|
# into ./output. Hand ownership back.
|
2026-06-11 07:46:48 +02:00
|
|
|
docker compose -f docker/docker-compose.yml exec kicad-wasm-builder \
|
|
|
|
|
chown -R "$(id -u):$(id -g)" /workspace/output || true
|
|
|
|
|
}
|
|
|
|
|
|
jspi: retire the asyncify pipeline — knob, post-link tail, binaryen hooks
Phase 8 in the parent repo. Deleted: asyncify-scheduler.js, apply-asyncify.sh,
apply-finalize.sh, inject-dyncall-shims.sh, asyncify-imports/removelist.txt,
the wasm-opt/finalize stub pair, scripts/binaryen-hoist-pass/ (the fork stays
a dormant submodule; removal is a follow-up), bench/wasm-opt-bench.sh (README
marked historical), wasm/shims/context_sleep.cpp, and the sched-context
harness app + Makefile targets.
PCBJAM_ASYNC_BACKEND is gone: build-wx-wasm.sh hardcodes the jspi stamp
(still force-cleans pre-migration trees), build-kicad-target.sh gives editors
the JSPI link surface and the CLIs nothing (they pin ASYNCIFY=0), the stub
dance is replaced by an unconditional .real-restore, build-wasm-test.sh lost
its whole post-link loop, docker/build.sh's postprocess is the ENV shim only,
and Makefile.wasm links every app JSPI with the scheduler shim as a tracked
prerequisite. pcbjam_async_policy.h keys on __EMSCRIPTEN__.
jspi-scheduler.js: wxWasmMainLoopPump dropped from the wrap census (the
export died with the D5 detach); inert [TRACE] instrumentation removed.
CI: wasm-build.yml rewritten for the single-cache pipeline (one output cache
keyed on compile inputs; post-processed bytes cached after the shim);
opt_level input removed from both callers. wasm-cache-hash.mjs inputs now
cover patch-env-shim.mjs + jspi-scheduler.js + jspi-exports.txt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 08:39:12 +02:00
|
|
|
# Phase 2 of one app: host-side post-processing (ENV merge shim). Pure host
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# work on output/${app}.js — no container needed.
|
2026-06-11 07:46:48 +02:00
|
|
|
postprocess_app() {
|
|
|
|
|
local app="$1"
|
2026-06-12 10:18:16 +02:00
|
|
|
local out_dir="output"
|
2026-06-11 07:46:48 +02:00
|
|
|
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# The headless CLI and the OCC/ngspice services skip the ENV merge shim:
|
|
|
|
|
# it exists for the interactive apps' runtime env overrides (?trace=),
|
|
|
|
|
# which these targets never read.
|
2026-07-19 15:59:21 +02:00
|
|
|
if [ "$app" = "kicad_tools" ] || [ "$app" = "occ_service" ] || [ "$app" = "ngspice_service" ]; then
|
2026-06-18 12:21:15 +02:00
|
|
|
echo "Skipping host post-processing for ${app} (finalized in-container)"
|
|
|
|
|
return 0
|
|
|
|
|
fi
|
|
|
|
|
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# ENV merge shim: the emscripten glue never merges Module.ENV into the
|
jspi: retire the asyncify pipeline — knob, post-link tail, binaryen hooks
Phase 8 in the parent repo. Deleted: asyncify-scheduler.js, apply-asyncify.sh,
apply-finalize.sh, inject-dyncall-shims.sh, asyncify-imports/removelist.txt,
the wasm-opt/finalize stub pair, scripts/binaryen-hoist-pass/ (the fork stays
a dormant submodule; removal is a follow-up), bench/wasm-opt-bench.sh (README
marked historical), wasm/shims/context_sleep.cpp, and the sched-context
harness app + Makefile targets.
PCBJAM_ASYNC_BACKEND is gone: build-wx-wasm.sh hardcodes the jspi stamp
(still force-cleans pre-migration trees), build-kicad-target.sh gives editors
the JSPI link surface and the CLIs nothing (they pin ASYNCIFY=0), the stub
dance is replaced by an unconditional .real-restore, build-wasm-test.sh lost
its whole post-link loop, docker/build.sh's postprocess is the ENV shim only,
and Makefile.wasm links every app JSPI with the scheduler shim as a tracked
prerequisite. pcbjam_async_policy.h keys on __EMSCRIPTEN__.
jspi-scheduler.js: wxWasmMainLoopPump dropped from the wrap census (the
export died with the D5 detach); inert [TRACE] instrumentation removed.
CI: wasm-build.yml rewritten for the single-cache pipeline (one output cache
keyed on compile inputs; post-processed bytes cached after the shim);
opt_level input removed from both callers. wasm-cache-hash.mjs inputs now
cover patch-env-shim.mjs + jspi-scheduler.js + jspi-exports.txt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NDeBaKKhQztd8KiVtHuyXr
2026-08-13 08:39:12 +02:00
|
|
|
# runtime ENV (?trace= would be a silent no-op — see docs/features/libs/0013).
|
|
|
|
|
kw_stage env-shim
|
2026-07-01 11:01:10 +02:00
|
|
|
node ./scripts/common/patch-env-shim.mjs "${out_dir}/${app}.js"
|
2026-05-29 06:39:23 +02:00
|
|
|
}
|
|
|
|
|
|
2026-06-11 07:46:48 +02:00
|
|
|
TOTAL_APPS="${#APPS[@]}"
|
2026-06-22 14:07:50 +02:00
|
|
|
|
|
|
|
|
if [[ "$PHASE" == "compile" ]]; then
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# --compile-only: produce the wasm+glue in output/; no host post-process.
|
2026-06-22 14:07:50 +02:00
|
|
|
idx=1
|
|
|
|
|
for app in "${APPS[@]}"; do
|
|
|
|
|
compile_app "$app" "$idx" "$TOTAL_APPS"
|
|
|
|
|
idx=$((idx + 1))
|
|
|
|
|
done
|
|
|
|
|
elif [[ "$PHASE" == "postprocess" ]]; then
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
# --postprocess-only: pure host post-process on the existing output/ glue
|
|
|
|
|
# (no container).
|
2026-06-11 07:46:48 +02:00
|
|
|
for app in "${APPS[@]}"; do
|
jspi cleanup: remove the asyncify-era residue — dead code, conditionals, pipeline scaffolding, stale prose
The runtime is JSPI-only; this removes everything that still pretended
otherwise. Three exhaustive sweeps (C++/JS+build+CI/tests+docs) drove
the inventory; every deletion verified by grep closure + full gates.
Broken-right-now fixes:
- deploy-staging.yml passed the retired opt_level input — the workflow
could not even start. Removed.
- env.sh carried dead exports with a live -sASYNCIFY=1 inside
(WASM_LDFLAGS/PTHREAD_LDFLAGS, zero consumers). Removed; the
WASM_LEGACY_EXCEPTIONS rationale rewritten to the real reason.
- docker/build.sh exported PCBJAM_ASYNC_BACKEND (read nowhere). Gone.
Dead weight removed:
- binaryen submodule (nothing builds or invokes it), wasm-opt-bench
workflow + scripts/bench/, get-wasm-opt.sh, diagnostics.js (242 lines
of Asyncify-API-only code), the KICAD_PIPELINE background-postprocess
scaffolding (existed to parallelize the deleted wasm-opt phase; the
postprocess is a seconds-long node script and now runs inline),
build-monitor's dead asyncify rows, sched-context orphan build
output, dead .gitignore entries, the .jspi-assets spike dir (the two
wf-result research JSONs moved to docs/features/async/migration-evidence/).
- bindings: fiber_park.h + its 12 embind registrations (broken-if-
called under JSPI), the kicadOpenFileStart/OPEN_JOB starter route,
main_stack_runner.h + 5 includes, the always-null context-sleep weak
hook in nanosleep_yield.c.
- shim: the backend field (installed-flag idempotency instead),
noteContextWait (dead both sides), the __wxAsyncifyDump alias (+ the
WasmTool fallback and string-dump normalize branch).
- web: the emscripten-6-ignored mainScriptUrlOrBlob option in boot.ts
(gerber-demo keeps it: it loads the deployed CDN release, which
predates emscripten 6 — noted inline).
Conditionals: all 'backend === jspi' checks reduced to scheduler-
presence checks; races_quiescent re-keyed from Asyncify.state (vacuous)
to real backlog quiescence (resumeReady/mutatorQueue — NOT _windowLive,
which is the probing activation's own window by definition).
Renames (identifiers only, no file renames): ASYNC_LINK_FLAGS→
JSPI_LINK_FLAGS and Makefile ASYNC_LDFLAGS→JSPI_LDFLAGS,
kicadCollabFiberBusy→kicadCollabBusy (embind + web + tests),
collab_common.h fiber*→apply*/coroutine naming, asyncifySignatures→
wasmTrapSignatures (lists byte-identical).
Tests: the two remaining vacuous [wx-asyncify]/fiber-resume-refused
asserts re-keyed to live JSPI beacons; eeschema-load's failure message
no longer sends the developer to a deleted script; wait-beacons' dead
families/parser deleted; lane-0 legacy-glue guards removed (lane 0 is
unconstructible); the embind test.fail re-gated with the JSPI reason
(plain embind invokers cannot suspend — verified still failing);
lint-determinism now scans tests/jspi (166 files clean);
eeschema-collab local-move gated to chromium (~50% flaky on FF even
solo; pcbnew twin covers both engines).
Docs: DEBUG.md rewritten as the JSPI debugging guide; build.md
describes the single-phase build; docs/features/async/README.md
banner-marked historical and repointed at the NEW
23-jspi-runtime.md (current architecture: export census, turnstile,
libcontext ownership + refusal contract, embind call shapes, the
em-pthread service-wrapper trick, exception policy, known gaps).
Gates on the cleaned tree: test:e2e 725 passed / 0 failed (after the
quiescence-probe fix; the 3 other reds were verified contention flakes
solo-green or the documented FF gate), web 76/0, jspi 18/18 both
engines, vitest 295/295 + 17/17, all lints green, live-app census
clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016X9eh1s5sTx1o9Em9KBuwR
2026-08-14 09:25:32 +02:00
|
|
|
postprocess_app "$app"
|
2026-06-11 07:46:48 +02:00
|
|
|
done
|
2026-05-29 06:39:23 +02:00
|
|
|
else
|
2026-06-22 14:07:50 +02:00
|
|
|
# both, sequential.
|
2026-06-11 07:46:48 +02:00
|
|
|
idx=1
|
|
|
|
|
for app in "${APPS[@]}"; do
|
|
|
|
|
compile_app "$app" "$idx" "$TOTAL_APPS"
|
|
|
|
|
postprocess_app "$app"
|
|
|
|
|
idx=$((idx + 1))
|
|
|
|
|
done
|
2026-05-29 06:39:23 +02:00
|
|
|
fi
|
2025-12-13 22:13:31 +01:00
|
|
|
|
|
|
|
|
echo ""
|
2025-12-08 12:07:01 +01:00
|
|
|
echo "Build complete. Output files in ./output/"
|
2025-12-13 22:13:31 +01:00
|
|
|
ls -lh output/
|