Host-side wxUniversal build + wx test apps (build-wxuniversal-wasm.sh /
build-wasm-test.sh, mirroring the dispatch-only ci.yml) followed by
`npm run test` before the KiCad e2e. The wx suite runs on bundled headless
Chromium; several specs use WebGL (gal-webgl.spec.ts), so the chromium
project gains a CI-gated --enable-unsafe-swiftshader launch arg — without
it newer headless Chromium refuses software WebGL on the GPU-less runner.
First time this suite runs on a working CI runner; treat as experimental.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The xvfb/WebGL fix (f37ddc4) took CI from 22 to 11 failures (identical on
Hetzner 27343415331 and Ubicloud 27343416511) — everything left is the
pcbnew family, still dying at wasm compile: pcbnew's ~190M debug module
exceeds SpiderMonkey's per-process code budget on x86-64 even with the
baseline-only JIT (the same module compiles on arm64, whose denser code
fits — which is why the local arm64 repro passed). No pref raises that
limit; V8 handles the module fine.
New CI-only 'chromium-ci' project (bundled Chromium, headless, SwiftShader
via --enable-unsafe-swiftshader) carries pcbnew/pcbnew-collab/load-pcb/
load-pcb-probe; the firefox project ignores them on CI. Workflows call the
new test:kicad:ci script which runs both projects. Local test:kicad is
unchanged (firefox everywhere, system Chrome for headed runs).
Validated under x86 emulation (linux/amd64 playwright image on Rosetta):
the pcbnew wizard test passes on chromium-ci in 23s against the fresh
v130-postprocessed 192M module.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
CI runs (Hetzner 27329612719, Ubicloud 27330989479) failed 22/41 identically,
in two buckets, both firefox-environment issues (the suite is firefox-only:
test:kicad -> --project=firefox):
1. No WebGL: headless Firefox can't create ANY GL context on a GPU-less VM
(blocklist bypass still dies with FEATURE_FAILURE_WEBGL_EXHAUSTED_DRIVERS),
so the GAL canvas failed and an error dialog blocked every UI flow. Fix:
run headed under xvfb (GLX + Mesa llvmpipe software GL) with
webgl.force-enabled — workflows now wrap the e2e step in xvfb-run.
2. pcbnew wasm OOM: the ~190M module OOMs Firefox's optimizing wasm JIT at
compile ("InternalError: out of memory") and the app never boots. Fix:
javascript.options.wasm_optimizingjit=false (baseline-only compile).
Both prefs are CI-gated in playwright-kicad.config.ts; local runs keep stock
behavior. Boot got slower (baseline JIT), so the wizard helpers now wait for
the registry to have actual UI ENTRIES (the registry OBJECT exists pre-boot)
instead of clicking into a fixed 10x500ms window.
Validated in a local docker repro of the CI env (playwright:v1.57.0 image,
no GPU): pl_editor canvas, gerbview wizard+canvas, and pcbnew wizard (on a
properly asyncified 187M module) all pass; before the fix they reproduced the
exact CI signatures.
Known-fail left on CI: zoom-cursor — the zoom-in anchors at ~ -P in headed
Firefox (screen-vs-client coordinate mix-up in the wheel->GAL path, headed
mode only); marked test.fixme(CI) pending a wx-wasm-layer investigation.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Suite went 64 failed → 291 passed / 0 failed:
- Makefile.wasm: compile with -MMD -MP and include .d files — stale objects
relinked against a newer wx lib crashed apps at startup with "function
signature mismatch" after any header/vtable change
- playwright(.coroutine).config.ts: pin the server port for the whole run
(resolvePort from playwright-kicad.config.ts); the 60s freshness window made
workers restarted after a failure rotate to a dead port (ERR_CONNECTION_REFUSED
cascade across ~50 tests)
- drop legacy-GL testing: KiCad renders via WebGL GAL (gal-webgl.spec.ts);
remove minimal_test's OpenGL tab/GLTestCanvas, opengl.spec.ts, GL z-order
describes, -sLEGACY_GL_EMULATION + gl_immediate_shim.js from the test build,
and the orphaned GL baselines
- coroutine-pthread repro apps join the `all` target (make clean used to delete
them while all never rebuilt them); clean no longer eats checked-in JS like
worker_dom_stub.js
- bump wxwidgets: guard module-eval document access so wx+pthread apps
(threadpool tests) survive Web Worker eval
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Reduce web/ to a single generic editor app plus a minimal example backend,
moving all project-specific/app code out to the closed root repo:
- apps/frontend -> standalone (@pcbjam/standalone): the GPL editor. Strips the
project-management UI (ProjectsPage/ProjectDetailPage/UploadDropzone and the
write half of lib/api.ts). WasmTool now takes fetchBytes + assetBaseUrl as
injected props (decoupled from the API client) so it can be driven by either
a backend or a local folder. New HomePage (local-folder loader + backend
project list) and read-only ProjectView. BroadcastChannel collab unchanged.
- backend (@pcbjam/backend-example): thin Fastify+ts-rest reference impl of the
@pcbjam/shared contract — serves a single project off the local filesystem
(PROJECT_DIR), no DB/auth/uploads.
- packages/contract -> web/pcbjam-shared (git submodule, MIT @pcbjam/shared).
- Remove apps/server, packages/storage, packages/contract, docker-compose
(server + storage move to the closed root repo; preserved via subtree
branches). Rewrite pnpm-workspace/turbo/.env for the frontend-only layout.
- Add tests/fixtures/demo so the editor + example backend run out of the box.
Standalone typechecks + builds; backend serves the contract end-to-end.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds a layer-preservation check to the footprint/via/zone/text round-trip add tests,
documenting + verifying that an item's own serialized layer is applied by makeItem and
the peer editor's active layer is never consulted. 4 add tests pass.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A board-level graphic text (Place→Text) lands in Drawings() and would otherwise hit the
same asyncify-fragile (kicad_pcb …) envelope-blob wall as via/zone on add. So PCB_TEXT
reconstructs NATIVELY: itemToJson emits size/thickness/angle plus horizontal & vertical
justification, mirror, and bold/italic (the text string was already emitted for any
EDA_TEXT); makeItem builds a fresh PCB_TEXT and restores all of them. flushDiff skips the
blob for PCB_TEXT_T. Footprint child text is unaffected — it syncs by move, and its add
is carried by the footprint blob.
Justification matters: it anchors the glyphs relative to the text POSITION, so without it
a left-justified text reconstructed centered on the peer and rendered visibly offset even
though GetPosition() matched (which is why the anchor-only headless check missed it).
Verified two-tab in the real app: a left/bottom-justified bold text added in tab A
reconstructs in tab B at the exact position with matching justification + string. Headless:
a board-text round-trip add test (sample text is left/bottom-justified) asserts the
position AND that hjust/vjust/text round-trip. 7 passed, 1 skipped, 0 aborts.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Footprint add reconstructs from the bare `(footprint …)` s-expr clipboard blob via
CLIPBOARD_IO (SetWriter/SetReader redirect it to a string for headless/wasm). The
non-footprint `(kicad_pcb …)` envelope parse (parseBOARD) is asyncify-fragile in wasm
— commit.Add of an envelope-parsed item silently stages nothing and subsequent virtual
dispatch traps ("index out of bounds"), the same wall that deferred the eeschema symbol
blob — so vias and zones reconstruct NATIVELY instead: itemToJson emits drill/layer-pair
for a PCB_VIA and the outline polygon for a ZONE, and makeItem builds a fresh PCB_VIA /
ZONE. flushDiff attaches the blob only to types that need it (footprints/board graphics);
tracks/vias/zones skip it. Includes the dangling-parent fix for the envelope path
(SetParent before delete) which footprints don't hit.
Build: kicad_clipboard.h pulls in the generated pcb_lexer.h (emitted into the common
build subdir by make_lexer) — added -I${KICAD_BUILD}/common to the embind include path.
Tests: sample board gains a via + zone; snapshot asserts their native fields are emitted,
and three round-trip add tests (footprint via blob, via/zone native) delete then re-add
each by uuid and confirm it returns at the same position. 6 passed, 1 skipped, 0 aborts.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Bug 1 — footprint child text now syncs. forEachTopItem also visits each footprint's
TEXT children by their own uuid (GetFields() PCB_FIELDs + GraphicalItems() PCB_TEXT),
so moving a silkscreen reference/value/user text — which leaves the footprint origin
unchanged — produces a diff. Apply needed only a guard: the removed loop skips items
with a parent footprint (a footprint delete cascades to its children, so don't
double-remove). ResolveItem already resolves child uuids and BOARD_COMMIT::undoLevelItem
rolls a child Modify up to the footprint, so Modify+SetPosition (in the existing
COROUTINE) just works. itemToJson now carries text for EDA_TEXT items.
Bug 2 — kicadCollabTestMoveFirst no longer no-ops. Its CallAfter now runs the move
inside a COROUTINE fiber (like doApply) so the virtual BOARD_ITEM::Move dispatches
instead of asyncify-no-opping.
Tests: sample board gains a footprint with a Reference field + user fp_text; snapshot
test asserts the footprint and both text children are emitted by uuid; a new test
moves each child by uuid and asserts the footprint origin stays put. 3 pass, 1 skipped.
Verified two-tab in the real app: C1's reference text move syncs A->B, no footprint moves.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
pcbnew's half of the unified Yjs collab bridge (4th tool; yjs-bridge commit 4),
a near-verbatim port of the eeschema design. Root-repo only — kicad/wxwidgets
submodules untouched.
- wasm/bindings/pcbnew_embind.cpp: BOARD_LISTENER trigger + post-settle snapshot
diff emit; BOARD_COMMIT apply inside a CallAfter + COROUTINE fiber (so a new
item's GAL view->Add dispatches correctly). Move/delete sync for any top-level
item by uuid; native PCB_TRACK add. Footprint/via/zone add deferred.
- WasmTool.tsx: add pcbnew to COLLAB_TOOLS.
- tests/apps/kicad/pcbnew-collab.html: seeded (wizard-free) harness, leaving
pcbnew.html untouched for its wizard test.
- tests/kicad/pcbnew-collab.spec.ts: snapshot + apply(move/remove/add) — 2 pass,
two-tab skipped headless.
Verified two-tab in the real web app: footprint move applies + syncs A->B.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Thrust A — dialogs render top-left with OK clipped in the WEB app (not the
test harness): root cause was the React shell missing the .window /
.window-canvas CSS that wx.js relies on (it positions each dialog div via
inline left/top, which need position:absolute). Added the rules to
web/apps/frontend/src/index.css. Native draw-text now works end-to-end;
symbol/power choosers render (placing still blocked by absent libraries).
Thrust B — collab apply of a newly-added SCH_SHAPE trapped in KiCad core
(SCH_COMMIT::Push CHT_ADD -> GAL view->Add, an asyncify invoke_* mis-dispatch)
because doApply ran off a fiber stack. doApply now runs inside a COROUTINE so
it executes on a libcontext fiber, the same context native draws use; the add
dispatches correctly. Re-enabled the SCH_SHAPE converter (rect/circle). Added
thirdparty/libcontext to the embind include path (tool/coroutine.h needs it).
Verified two-tab: rectangle + circle drawn in tab A sync + render in tab B.
Extended eeschema-collab.spec.ts apply test with a SCH_SHAPE add.
All changes root-repo only; kicad and wxwidgets forks untouched.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Bump kicad + wxwidgets submodules with two wasm eeschema UI fixes, and add the
regression test tests/kicad/eeschema-ui.spec.ts (Delete + Backspace delete; the
text-tool properties dialog opens and closes without freezing).
- wxwidgets c27fe8b: nested (quasi-modal) event loops pump via Asyncify instead of
re-entering emscripten_set_main_loop (which threw an un-resumable 'unwind').
- kicad 4132395: bind Backspace to delete under emscripten + apply default alt hotkeys.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Two things, both verified in the real web app (two-tab eeschema collab).
1. dynCall crash fix (all apps) — scripts/common/shims/dyncall-binding.js.tmpl.
Programmatic editor edits trapped with 'indirect call signature mismatch': the
asyncify-instrumented wasmExports[dynCall_<sig>] trampoline does call_indirect with a
stale type for some table indices (post-asyncify+O2) even though the table entry is
valid. Proven by patching the built js: at the trap getWasmTableEntry(index) SUCCEEDS
where the trampoline fails. Fix: the shim now catches the 'signature mismatch'
RuntimeError and falls back to getWasmTableEntry; the Asyncify unwind sentinel and real
exceptions re-throw, so instrumentation/unwind is untouched for normal calls. This
unblocks ALL programmatic edits, not just collab (e.g. eeschema SCH_ITEM::Move).
2. eeschema collab apply converters (wasm/bindings/eeschema_embind.cpp).
doApply now handles added-item construction (build the SCH_ITEM with the delta's uuid
via const_cast — as the s-expr parser does — + commit.Add) and richer SCH_LINE
serialization (start/end/layer) so wire edits reconstruct on the peer. Implemented for
SCH_LINE (wires) + SCH_JUNCTION; other types log 'no converter for added type' and are
skipped (next batch). eeschema re-enabled in the web app collab gate.
Tests: eeschema-collab.spec snapshot (green); apply/two-tab skipped — they no-op headless
because the e2e harness's kicadOpenFile returns false (OpenProjectFiles bails before
building the connectivity graph), so SCH_COMMIT::Push doesn't persist. Verified in-app.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
eeschema's half of the Yjs collaborative bridge, reusing the generic reconciler /
BroadcastChannel transport unchanged. Zero kicad-fork change: native SCH_ITEM uuid +
native SCHEMATIC_LISTENER. All in the wasm layer (wasm/bindings/eeschema_embind.cpp).
Working (verified in the web app):
- kicadCollabSnapshot(): enumerate sch.Hierarchy() -> LastScreen()->Items() as
{id,type,x,y}; registers the listener on first call
- emit: SCHEMATIC_LISTENER subclass -> per-item delta via window.kicadCollab.onDelta;
fires on real SCH_COMMIT::Push (a real wire move broadcasts added/removed/changed)
Apply is a documented follow-up (gated off so a peer tab can't crash): SCH_ITEM::Move
traps with 'indirect call signature mismatch' when invoked outside a KiCad tool
coroutine (Asyncify+fiber+exception-trampoline). Modify/Clone/GetPosition all work;
only the virtual Move write traps. Fix direction: route apply through TOOL_MANAGER.
Also: build-kicad-target.sh now force-relinks when only <app>_embind.cpp changed (the
embind .o isn't a make dep, so new bindings silently vanished), and adds the
expected/rtree/fmt thirdparty includes the eeschema bindings need.
Tests: eeschema-collab.spec.ts covers snapshot (green); apply/two-tab skipped with the
blocker noted. WasmTool gates collab to pl_editor only until eeschema apply works.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Bump kicad submodule to the per-item uuid identity change, plus the wasm/test
infra to verify it:
- wasm/bindings/pl_editor_embind.cpp: test-only kicadSaveDrawingSheet(path) hook
that serializes the singleton DS_DATA_MODEL to MEMFS (also a building block for
the bridge's later materialize-to-file path)
- tests/kicad/pl_editor-uuid.spec.ts: open->save->read-back e2e proving (uuid …)
backfill (4 distinct uuids) and load->save round-trip preservation
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Register the newly WASM-ported Gerber Viewer in the web app the same way as
the other tools — registry entries only, no UI edits (ProjectDetailPage renders
launch links generically from FILELESS_TOOLS).
GERBVIEW_FRAME opens gerber/drill files through its own File→Open UI and projects
carry no gerber files to auto-open, so it's treated as file-less (boot standalone),
mirroring symbol_editor. It boots through single_top.cpp's STARTWIZARD, so it seeds
config to skip the first-run wizard (TOOL_NEEDS_CONFIG_SEED) and gets a
/usr/bin/gerbview argv0.
- contract: add "gerbview" to TOOLS, TOOL_LABELS ("Gerber Viewer"), FILELESS_TOOLS
- frontend: add gerbview to TOOL_ARGV0 and TOOL_NEEDS_CONFIG_SEED
- e2e: add file-less gerbview case to tools-open.spec.ts (title "Gerber Viewer",
canvas painted, wizard-free, no WASM abort)
Verified in-browser via npm run test:web: gerbview boots wizard-free with the
viewer chrome (toolbars + layers manager), 0 console errors.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add an e2e suite that drives the real React web app (not the standalone
harness): tests/playwright-web.config.ts + tests/web/tools-open.spec.ts navigate
/p/demo/<tool>/<file> for all five tools and assert each boots, opens its demo
file (title drops "untitled"), shows no first-run wizard, and emits no WASM abort
or URL-regex modal. global-setup-web.ts re-seeds the demo project through the API
if missing, so the suite is self-sufficient against a running dev stack. Wired as
`npm run test:web`.
Also at the harness level:
- pl_editor-load.spec.ts: prove the pl_editor kicadOpenFile embind hook opens a
.kicad_wks (mirrors eeschema-load.spec.ts).
- seed KiCad config in pl_editor.html / symbol_editor.html (matching eeschema.html
and the web app's boot.ts) so the harness boots wizard-free; repurpose
pl_editor.spec.ts's stale "wizard completes" test into a wizard-skip regression.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add tests/kicad/eeschema-url-regex.spec.ts: opens a text_box-bearing
schematic with a real URL via Module.kicadOpenFile and asserts the
URL-detection wxRegEx compile no longer fails — no "Invalid regular
expression" in the console and no error dialog. A pre-fix build renders the
same text_box, hits IsURL() -> the failing static regex, and fires both
signals.
Bumps the wxwidgets submodule to the wxConvLibc->UTF-8 emscripten fix.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The emscripten fiber glue gates Fibers.trampoline() on Fibers.trampolineRunning
and resets it at the end of its loop. At startup emscripten_set_main_loop(...,1)
throws "unwind" to establish the main loop, and KiCad does so from inside a tool
coroutine, so the throw propagates THROUGH the trampoline and skips the reset —
leaving the flag stuck true. Every fiber swap after startup then becomes a silent
no-op, so opening a schematic (SetScreen -> RunAction(selectionClear) -> fiber
swap) hangs forever with the editor stuck on "untitled".
Wrap the trampoline loop in try/finally (inject-dyncall-shims.sh section "3c") so
the flag is always reset. Add tests/kicad/eeschema-load.spec.ts, which opens a
small wires/junctions schematic via Module.kicadOpenFile and asserts the editor
title switches away from "untitled": it times out (RED) without the shim and
passes (GREEN) with it. Also add features/web-init/0002-url-regex-modal-followup.md
capturing the unrelated URL-detection wxRegEx modal surfaced once loading works.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds a 600ms wait between the wxFileDialog appearing in the registry and
the screenshot in tests/kicad/pl_editor.spec.ts. Without it, the dialog
object exists but its MEMFS readdir (asyncified) hasn't returned yet, so
the screenshot caught a half-painted black rectangle — fine for the
registry-based assertion but useless as a pixel baseline.
Commits the 10 pl_editor screenshots produced by the spec as the baseline
reference under tests/baseline-screenshots/.
Bumps the wxwidgets submodule pointer to pick up the expanded comment on
the wxGenericFileDialog::OnOk fix (explains why the patch lives in the
generic dialog instead of src/wasm/ — OnOk is the join point for both the
OK button and <Enter> via the dialog's compile-time event table, so
subclassing in wasm/ wouldn't intercept either).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Brings up KiCad's pagelayout_editor (drawing-sheet editor) in the
browser, to roughly the same "boots, canvas visible, partially usable
in-session" level as the existing pcbnew/eeschema/calculator ports.
Build:
- docker/build.sh: add pl_editor to the unified app dispatch (case,
subdir map, all-loop).
- scripts/kicad/build-kicad-target.sh: add pl_editor to the case;
upstream target name pl_editor under source subdir pagelayout_editor.
- scripts/kicad/build-pl_editor.sh: 7-line thin wrapper matching the
pcbnew/eeschema/calculator pattern.
- tests/scripts/setup-kicad-wasm.sh: copy_app pl_editor.
App glue:
- wasm/stubs/nl_pl_editor_plugin_stub.cpp: no-op SpaceMouse plugin so
pl_editor_frame.cpp's NL_PL_EDITOR_PLUGIN symbols resolve. Mirrors
nl_pcbnew_plugin_stub.cpp.
- tests/apps/kicad/pl_editor.html: browser shell. preRun creates
/home/kicad and FS.chdir there so file dialogs land somewhere
friendly instead of MEMFS root (/dev/, /proc/, etc.).
E2E coverage:
- tests/kicad/pl_editor.spec.ts: 5 tests — smoke (canvas, no abort),
wizard, File menu has Open/Save As, file-dialog folder-navigation
regression, canvas + toolbar metrics.
- tests/e2e/filedialog-folder-nav.spec.ts: wxWidgets-level twin of
the regression test (exercises the underlying widget directly via
the standalone filedialog_test app).
Submodule bumps:
- kicad → feature/pl-editor (WASM gating in pagelayout_editor's
CMakeLists + navlib stub).
- wxwidgets → feature/pl-editor (wxGenericFileDialog::OnOk navigates
into selected directories; wasm/mouse.cpp emits wxEVT_LEFT_DCLICK
via timestamp-based double-click detection — the latter benefits
every wxWidgets-WASM app).
See features/pl-editor/ for the design doc + per-repo diff patches.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Convert the calculator port (commit 0464470) from the parallel/copy-paste
pattern to the dispatch pattern used by pcbnew + eeschema.
- docker/build.sh: <app> is now required (no pcbnew default); missing,
unknown, and --help paths all print "pcbnew | eeschema | calculator | all".
Adds calculator to build_app() via kicad_subdir_for() (pcb_calculator
inner subdir, calculator.* output basename). Also fixes the
EMSDK=/emsdkkicad-wasm-builder typo that snuck in during the calc merge.
- scripts/kicad/build-kicad-target.sh: accepts calculator; introduces
KICAD_TARGET (pcb_calculator for calc, == APP_NAME otherwise) used for
the make target, embind include path, and final-log line.
- scripts/kicad/build-calculator.sh: 305-line copy of build-pcbnew.sh
collapsed to a 7-line wrapper around build-kicad-target.sh.
- tests/scripts/setup-kicad-wasm.sh: copy_app calculator added with the
same pcb_calculator subdir mapping for the docker-volume fallback path.
- tests/package.json: test:calculator* routed through the shared
playwright-kicad.config.ts kicad/calculator.spec.ts (mirrors eeschema).
- Delete: docker/build-calculator.sh, tests/scripts/setup-calculator-wasm.sh,
tests/playwright-calculator.config.ts.
Bumps wxwidgets d1d1627 -> 6fb2eac (origin/wasm-port). The new sha includes
"unic/combobox: add GetCurrentSelection() inline default" which calc's
kicad/pcb_calculator/widgets/unit_selector.cpp needs to compile. Without
this bump the unified dispatch would expose calc as a buildable target but
the build itself would fail. Verified: build.sh all completes clean across
all three apps; pcbnew/eeschema e2e pass; calc compiles and launches (test
stability separate from this refactor).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Adds an end-to-end test that drives File→Open in pcbnew, injects
the .kicad_pcb and .kicad_pro files into MEMFS at the dialog's
default starting directory, drives the menu + filename text input
+ Enter accept path, and screenshots the loaded board. Parametrized
for both kicad/demos/microwave (RF polygon footprints) and
kicad/demos/pic_programmer (full multi-IC layout).
Without the rtree fix bumped in via the kicad submodule, the load
would abort on every PCB at rtree.h:1771 Classify; the test asserts
no [RTREE-DIAG] line and no Aborted(. The post-load clipboard
RuntimeError in __asyncjs__js_clipboardHasText is a separate,
pre-existing wasm-port limitation that we explicitly do not regress
on here.
- tests/kicad/load-pcb.spec.ts: serial-mode parametrized spec
- tests/kicad/load-pcb-probe.spec.ts: one-shot diagnostic probe
for inspecting wxFileDialog state on the canvas
- tests/kicad/utils/fs-inject.ts: FS.writeFile bridge from Node fs
- tests/kicad/utils/board-ready.ts: poll-for-no-dialogs readiness
- tests/baseline-screenshots/load-pcb-*.png: 6 baselines covering
both demos at pcbnew-ready / dialog-open / loaded states
- features/.../rtree-debug-findings.md: full diagnosis trail with
an upstream-reportable summary the maintainer can lift verbatim
- kicad submodule bumped to 07d8130d44 (shape_poly_set rtree fix)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
- build-pcbnew.sh: add --diag=<gal,coroutine,ctor,all> -> -DKICAD_DIAG_*,
off by default (forwarded by docker/build.sh)
- diagnostics.js: emit at console.log level (no longer error/warn); still
gated by SHIM_DIAGNOSTICS=1
- apply-asyncify.sh: exclude PCB_EDIT_FRAME::setupUIConditions() from
asyncify instrumentation (V8 cannot run the instrumented huge function
on the rewound ctor stack -> Chrome startup stall; Firefox unaffected)
- DEBUG.md: reusable WASM/asyncify/browser debugging guide, diagnostic
flag docs, and a production-build (release + -O2 asyncify) recipe
- tests: standalone coroutine vcall/gl repro probes
- bump kicad + wxwidgets submodules (diagnostic gating / debug cleanup)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Investigation scaffolding for the Chrome-only KiCad coroutine renderer crash.
Adds isolated reproduction probes exercising the coroutine/Asyncify/fiber layer
under KiCad-like conditions, runnable in BOTH Firefox and system Chrome.
- tests/playwright-coroutine.config.ts + test:coroutine:firefox|chrome npm
scripts: run the coroutine specs in Firefox AND system Chrome (the old e2e
config only used bundled Chromium, which never reproduced the crash).
- tests/apps/standalone/coroutine-pthread/: no-wx + pthreads reproduction probes
(fiber-in-main, nested invoke_/dynCall boundaries, RunMainStack, embind,
main-loop/rAF activation) + worker_dom_stub.js for wx+pthreads builds.
- tests/apps/Makefile.wasm: coroutine-pthread{,-main,-nested,-nested-ex,-wx,
-embind,-mainloop} targets.
- scripts/common/shims/diagnostics.js: add EM_ASYNC_JS handleSleep enter/wake
tracking (DIAG_SLEEP) to detect nested-async at the crash.
Findings (details in research notes): every isolated factor so far — direct /
nested / RunMainStack fiber, wx event loop + all 13 scenarios incl EM_ASYNC_JS,
pthreads, and main-loop/rAF activation — runs CLEAN in system Chrome. The
coroutine/Asyncify layer is exonerated; GL/WebGL is the remaining untested factor
(next). The reliable FF-pass/Chrome-fail repro is still the KiCad pcbnew e2e.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The shim bound bare dynCall_* names to JS getWasmTableEntry() calls, bypassing
the asyncify-instrumented dynCall_* wasm trampolines that -sDYNCALLS=1 provides.
That broke Asyncify unwind/rewind through indirect calls -> "indirect call
signature mismatch" (caught every frame in Firefox; fatal renderer crash in
Chrome). Bind the bare names to wasmExports["dynCall_<sig>"] instead.
Result: the PCBnew "select draw lines" e2e is green in Firefox (tool selects and
draws, zero page errors). Dropped the fiber-stabilization block, the shipped
diagnostic block, and the exportCallStack JS hack (all compensated for the wrong
binding); shim shrank 521 -> ~250 lines.
- scripts/common/inject-dyncall-shims.sh: orchestrator only; injected JS extracted
to scripts/common/shims/
- scripts/common/shims/dyncall-binding.js.tmpl: per-signature binding template
- scripts/common/shims/handlesleep.js: nested-Asyncify handleSleep fix (#9153)
- scripts/common/shims/diagnostics.js: logging-only, opt-in via SHIM_DIAGNOSTICS=1
- tests/package.json: add test:kicad:firefox / test:kicad:chrome scripts
Known issue (tracked separately): Chrome still renderer-crashes on the first
coroutine resume. Asyncify.doRewind replays the deep main-context call stack and
exceeds V8's execution-stack limit (Firefox tolerates the same wasm). Proper fix
is JSPI.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Main-repo side of a multi-part WIP covering the KiCad WASM tool-selection
and nested-Asyncify work. Submodule commits are in kicad@f6e9239aaa
(libcontext hygiene) and wxwidgets@bb80f91e8b (auibar registration +
dialog diagnostics).
## scripts/common/inject-dyncall-shims.sh
Wrap Asyncify.handleSleep / allocateData to save-and-restore Asyncify.currData
around each EM_ASYNC_JS sleep. This fixes the nested Asyncify collision where
a fiber swap that fired during a modal's event loop clobbered currData, and
the modal's later doRewind used the fiber's buffer and hit "RuntimeError:
index out of bounds". Root cause documented as Emscripten Issue #9153
(wontfix upstream).
Diagnostic-rewind logging (forcedBottomOfCallStack, callStack traces) is
retained to help future debugging of Asyncify state corruption.
## tests/
- tests/playwright-kicad.config.ts: add `channel: 'chrome'` for the
chromium project so --project=chromium --headed uses system Chrome
(real GPU) instead of SwiftShader on ARM Mac. Also switch trace to
retain-on-failure + screenshot on-failure for easier E2E debugging.
- tests/kicad/pcbnew.spec.ts: replace `tool.checked` assertions with a
label-suffix check (`[checked]`) since our auibar registration encodes
checked state in the label (no schema change to the registry).
- tests/apps/Makefile.wasm: add `coroutine-nested` build target + include
it in the all: list.
- tests/apps/standalone/coroutine/: kicad_coroutine_harness.h + test app
reproducing KiCad COROUTINE semantics against real libcontext.
- tests/apps/standalone/coroutine-nested/: nested_test.cpp reproduces the
EM_ASYNC_JS-modal + fiber-swap nesting bug in isolation. 8 scenarios
from baseline_modal_alone through nested_fibers_inside_modal.
- tests/e2e/coroutine.spec.ts + coroutine-nested.spec.ts: Playwright specs
that load the standalone apps and assert all case cases pass via
[COROUTINE_TEST] SUMMARY log parsing.
## research/ and features/browser-tools/
Three background docs capturing the investigation trajectory:
- features/browser-tools/0001-kicad-wasm-tool-activation-investigation.md
Early investigation: why tools don't activate; initial dynCall-empty-
callback hypothesis.
- features/browser-tools/0002-wasm-coroutine-deep-dive.md
Deep dive on Asyncify internals, fiber API, QEMU's coroutine-wasm
reference implementation.
- features/browser-tools/0003-wxauitoolbar-registration-fix.md
The narrow fix: why wxAuiToolBar needs a registration block, where to
add it, what the fallback plan is.
- research/threading_1.md: corrected root-cause analysis after reading
runtime logs — nested-Asyncify currData collision, Emscripten #9153.
- research/threading_2.md: extended research on alternative approaches
(JSPI/WasmFX/state-machines) and why they don't help here.
## Submodule pointer updates
kicad: f6e9239aaa (wip: libcontext WASM hygiene cleanup)
wxwidgets: bb80f91e8b (wip: wxAuiToolBar element-registry registration +
dialog diagnostics)
## Open threads not yet in scope
- Firefox/Chrome divergent behavior: "indirect call signature mismatch"
traps in Firefox vs renderer crash in system Chrome (tracked in
plans/peaceful-hugging-pnueli.md and the research docs).
- E2E pixel-diff for Draw Lines fails because the test's diff region does
not cover where the line is actually drawn; tool activation works, the
line is visible in test-results/pcbnew-draw-lines-02-after-drawing.png.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Update wxwidgets submodule with fix for consecutive modal dialog crashes.
The fix prevents overlapping Asyncify operations that caused "indirect call
to null" and "func is not a function" errors when a second modal was
triggered immediately after the first one completed.
Also includes:
- docs: Clarify build script order and descriptions in CLAUDE.md
- refactor(test): Remove debug logging from wizard test
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Test now correctly finds "Finish" button on last wizard page
- Added CLAUDE.md note about running e2e tests via npm scripts
- Simplified test to click through wizard with proper button detection
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
README updates:
- Document WebGL GAL integration in kicad/common/gal/webgl/
- Add Test Scripts section with all 4 GAL scripts
- Add WebGL Integration section explaining the architecture
- Update Directory Structure to include wasm/ and baseline-webgl/
- Add Comparison Thresholds section
Baseline updates:
- Refresh 20 WebGL baseline images after KiCad integration
- Minor anti-aliasing differences from previous baselines
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Migrate WebGL GAL implementation from tests/gal-regression/wasm/webgl/
to kicad/common/gal/webgl/ and kicad/include/gal/webgl/.
This integrates the WebGL GAL properly into KiCad's build system:
- Update test Makefile to use sources from kicad/ instead of local copies
- Update build scripts for new source locations
- Add test-gal-webgl.sh script for running WebGL regression tests
- Update Docker to Emscripten 4.0.22
The WebGL GAL passes all 28 regression tests (matching baseline).
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Establish baseline screenshots for WebGL GAL implementation before
migrating it from test harness into KiCad's build system.
Current status: 7/27 scenarios passing (20 different from native):
- Alpha-blending: FIXED (blending now correct)
- Transform-API: EXCLUDED (dead code)
- Other scenarios: Work in progress
This baseline will be used by test-gal-webgl.sh to detect regressions
during the migration process. Goal is to preserve current rendering
fidelity while integrating WebGL GAL into KiCad source tree.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Changes:
- Exclude gal-transform-api.png from native/WebGL comparison since
Transform() is dead code in KiCad (never called, has no effect)
- Add verbose diagnostic output showing pixel difference details,
content bounds, and sample pixels for failing scenarios
- Update README to document that Transform() is dead code in both
native and WebGL implementations
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Alpha-blended shapes were rendering nearly invisible due to incorrect
blend function configuration in two places:
1. FBO rendering (webgl_gal.cpp): Changed to glBlendFuncSeparate to handle
RGB and alpha channels independently. Alpha channel now accumulates
coverage correctly (prevents it from staying near 0.0 when rendering
with alpha=0.5).
2. Compositor (webgl_compositor.cpp): Changed from premultiplied alpha
blend (GL_ONE, GL_ONE_MINUS_SRC_ALPHA) to straight alpha blend
(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA) since our FBOs use straight alpha.
Also marked Transform() as dead code with explanation - never called in
KiCad and has no effect even in native OPENGL_GAL.
Result: Alpha-blending scenario now renders correctly with proper color
mixing for overlapping shapes.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Native OpenGL screenshots have transparent backgrounds (alpha=0 in undrawn
areas). The comparison script flattens both images to white before comparing.
WebGL must use white clear color to match this behavior.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
