pcbjam/tests/kicad/load-pcb.spec.ts
Istvan Matejcsok c421d724b0 findings(E-10..E-22): fix the defects a code review found in the E-1..E-9 work
A review of the group-E fixes found 13 further defects; ten were introduced by
those fixes, two pre-existed and were merely relocated, one is deferred.

Services / transport
  E-10  retireWorker synthesized no bg/exit frame, so sharedspice's s_bgRunning
        mirror stayed latched true after a mid-run worker death: Run stayed
        disabled and the promised fresh-worker restart was unreachable for the
        whole session. Retirement now dispatches a synthetic controlled-exit
        straight to the installed handler (never through dispatchEvt — a
        fabricated frame must not touch the credit ledger). Driving the repro
        exposed two further defects, both fixed here: a replacement worker
        trapped on pre-init engine reads, and the rerun's cm_input_path/circ hit
        that uninitialized engine before KiCad's validate() re-init (the native
        flow assumes a crashed engine survives in-process — true for the dll,
        false for a dead worker). Reads now answer their empty shapes pre-init,
        writes lazy-init, and init is idempotent per worker engine.
  E-19  dispatchEvt acked only AFTER handler(evt) returned, and the sharedspice
        client deliberately rethrows non-trap errors — so each throw leaked one
        unit of the 64-frame credit window until the stream died with a
        misattributed "transport exceeded". The ack moves to a finally in both
        service copies; the throw still propagates (the trap machinery needs it).
  E-20  the oversize-line path promises to transfer the accepted prefix, but
        with the window full that flush only DEFERS, and stopEventStream wiped
        the deferred queue — losing the diagnostics that explain the failure.
        The terminal notice now carries them as pendingEvents; both hosts
        deliver them in order, unacked (the fatal frame is outside the credit
        protocol).
  E-21  the 30s prefetch deadline discarded every model already collected and
        reported nothing. A caller-owned progress sink ships the partials and
        the omission reaches the export report. (Awaiting the aborted collection
        was rejected: an in-flight source fetch is not abortable — E-4's
        original disease.) Plus a serving-candidate memo, so a .wrl ref served
        by its .step fallback stops re-probing the miss on every export.

Scheduler
  E-14  _terminalizeNativeTrap classified by message substring, so any plain JS
        error QUOTING 'Aborted(' or 'out of bounds' permanently bricked a
        healthy instance. Now structural only: instanceof RuntimeError plus a
        duck-typed name check (verified in this build's glue that abort() throws
        a genuine RuntimeError both pre- and post-runtime-init). Module.onAbort
        now latches the gate — the authoritative notification, previously
        ignored.
  E-15  the shim half: _pumpResume gates on terminal (catching wakes already
        queued at latch time) and resolveWait refuses on terminal WITHOUT
        consuming the entry, so a frame stays visibly parked rather than
        resuming inside a trapped module.
  E-16  the E-5 handler read the realm-global scheduler at dispatch instead of
        its installing module's; also frees the per-line buffer on the non-trap
        rethrow path.
  E-11  get_vec trusted the worker's res.length over the transferred arrays.
        Observed death shape: a 4 GiB std::vector threw an unhandled
        std::length_error that exited the editor's main loop. Now clamped, with
        the buffers freed on every failure path.

Guardrails (replacing two deferred refactors: e2e→production-code injection and
collapsing the four copies of the worker-lifecycle machinery)
  E-18  the source contract asserted comment-string counts — rewording failed
        CI while moving a guard outside its #ifdef passed. It now parses the
        #ifdef regions and asserts on code.
        service-stub-parity.ts pins what the four lifecycle copies must share:
        credit-window equality parsed from source, the finally-ack, boot
        deadlines, terminal-notice consumption. The transport numbers are now
        single-sourced from the worker.
        CI actually runs the gates: the web/standalone vitest suites (which had
        NEVER run in CI), the reducer, the source contract and the parity tool —
        with a NON_PLAYWRIGHT_GATES check so deleting a step re-fails the lint.
  E-22  the e2e occ stub's 60s boot watchdog, deleted in a66e109, is restored in
        the ngspice-stub shape with a wedgeNextBoot() repro hook.

Every behavioral fix has red-then-green evidence (the reds were captured first).
E-17 (a stale RUNNING cross-stamping the next run's generation under E-6's
transport deferral) is DEFERRED with its analysis recorded — a real fix needs
run identity on the bg frames.

Test hygiene: the dwell lint now requires the mandated ": <why>" and all 47 bare
markers carry their reason; three export-report dwells became modal-lease polls;
exact-ledger assertions became relative deltas; the dead data-wx-dom-id branch,
an unused fault hook and unused receipt plumbing are gone; abort scans, wx
dialog drivers, the sim harness and the vitest FakeWorker are each one copy now.

Bumps kicad and wxwidgets to their findings-group-e tips.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-08-31 18:19:16 +02:00

215 lines
11 KiB
TypeScript

import type { Page } from '@playwright/test';
import { test, expect } from './fixtures';
import {
clickMenuBarItem,
clickMenuItem,
waitForEditorReady,
waitForRenderedByLabel,
waitUntil, stableShot, shotPath } from '../e2e/utils/element-tracker';
import { injectFromSubmodule } from './utils/fs-inject';
import { waitForBoardLoaded } from './utils/board-ready';
/**
* Spike test: load real .kicad_pcb demos through pcbnew's File → Open.
*
* Flow (parametrized for each demo below):
* 1. Wait for pcbnew to come up (dismiss the first-run KiCad Setup wizard
* if it appears — depends on whether MEMFS already has config).
* 2. Inject the demo's .kicad_pcb (+ .kicad_pro for completeness) into
* MEMFS at PATHS::GetDefaultUserProjectsPath() — confirmed by
* load-pcb-probe.spec.ts to be /home/kicad/documents/kicad/10.0/projects/.
* 3. Drive File → Open via menu helpers.
* 4. Click on the file row in wxFileListCtrl (wasm port doesn't register
* listctrl rows in wxElementRegistry, so we click by the filelist's
* bounding box), then focus the filename text input and press Enter to
* trigger wxGenericFileDialog's accept path. (A direct OK button click
* doesn't dismiss the dialog in the wasm port; the listctrl row click
* populates the text input but doesn't mark the row selected enough for
* OK to satisfy validation.)
* 5. Dismiss any post-load info dialogs (missing-libs etc. — pic_programmer
* uses local footprint libs and may pop one of these).
* 6. Wait until the file dialog and the LoadBoard progress dialog both go
* away. Screenshot the loaded board, named per-demo.
*
* The board load itself was crashing in `Classify` inside `rtree.h` until we
* fixed the wasm-only `ELEMTYPEREAL = intptr_t` (32-bit on wasm32) typo at
* `kicad/libs/kimath/src/geometry/shape_poly_set.cpp:1927`. See
* `features/<branch>/rtree-debug-findings.md` for the full trail.
*/
// KiCad 10 stores projects under /home/kicad/documents/kicad/10.0/projects
// (GetMajorMinorVersion() = "10.0"); the old WASM port used "9.99".
const KICAD_VERSION_DIR = '10.0';
const PROJECT_DIR_MEMFS = `/home/kicad/documents/kicad/${KICAD_VERSION_DIR}/projects`;
type DemoCfg = {
name: string; // appears in test name & screenshot filename
dir: string; // folder under kicad/demos/
stem: string; // file stem (kicad_pcb / kicad_pro)
};
const DEMOS: DemoCfg[] = [
{
// RF-polygon-heavy board — small, no external libs, was the case that
// tripped the rtree overflow in the first place.
name: 'microwave',
dir: 'microwave',
stem: 'microwave',
},
{
// "Normal" through-hole/SMD design — larger (~698 KB), uses local
// footprint libs from its own project tree (may pop a missing-libs
// warning, which we dismiss before screenshotting).
name: 'pic_programmer',
dir: 'pic_programmer',
stem: 'pic_programmer',
},
];
function runLoadPcbTest(demo: DemoCfg): void {
const pcbFilename = `${demo.stem}.kicad_pcb`;
const proFilename = `${demo.stem}.kicad_pro`;
test(`opens ${demo.name} demo from MEMFS through the wxFileDialog`, async ({ page, testLogger }) => {
await page.goto('/kicad/pcbnew.html');
await waitForEditorReady(page);
await stableShot(page, `load-pcb-${demo.name}-00-pcbnew-ready.png`);
// ── Inject .kicad_pcb + .kicad_pro into the dialog's start dir. ──
await injectFromSubmodule(
page,
`kicad/demos/${demo.dir}/${pcbFilename}`,
`${PROJECT_DIR_MEMFS}/${pcbFilename}`,
);
await injectFromSubmodule(
page,
`kicad/demos/${demo.dir}/${proFilename}`,
`${PROJECT_DIR_MEMFS}/${proFilename}`,
);
// ── Drive the menu. ────────────────────────────────────────────
const fileClicked = await clickMenuBarItem(page, 'File');
expect(fileClicked, 'File menu should be findable').toBe(true);
await waitUntil(
page,
() => {
const r = window.wxElementRegistry;
if (!r?.findAllRendered) return false;
return r.findAllRendered({ elementType: 'menuitem' }).length > 3;
},
'File menu items rendered',
);
// Items register progressively while the popup paints — wait for the one
// we click (clickMenuItem is single-shot; the >3-items gate isn't enough).
await waitForRenderedByLabel(page, 'Open...', { elementType: 'menuitem' });
const openClicked = await clickMenuItem(page, 'Open...');
expect(openClicked, 'Open… menu item should be findable').toBe(true);
// ── Wait for the wxFileDialog to appear, file list to paint. ───
await page.waitForFunction(() => {
const registry = window.wxElementRegistry;
if (!registry) return false;
return registry.findAll({ visible: true })
.some((el) => el.typeName === 'wxFileDialog');
}, null, { timeout: 15000 });
// stableShot stabilizes the file-list paint before comparing — deterministically
// replacing a fixed 1000ms that used to catch the dialog mid-paint (black rectangle).
await stableShot(page, `load-pcb-${demo.name}-01-dialog-open.png`);
// ── Focus the filename text field, type the name, accept. ──────
// The Open dialog gives default keyboard focus to the file LIST,
// where keystrokes act as type-ahead (and Enter on the highlighted
// ".." row navigates up) rather than filename entry. So we must click
// the wxTextCtrl to focus it first, located via the element registry
// rather than a fixed pixel offset. No filelist row click is needed —
// the earlier version's extra row + offset clicks were what left focus
// off the field, so the typed name never registered, the field stayed
// empty, and Enter was a no-op (the board never loaded).
const filenameInput = await page.evaluate(() => {
const registry = window.wxElementRegistry;
if (!registry) return null;
const text = registry.findAll({ visible: true })
.find((el) => el.typeName === 'wxTextCtrl' && el.name === 'text');
return text ? { x: text.centerX, y: text.centerY } : null;
});
expect(filenameInput, 'filename text input should be visible').not.toBeNull();
if (!filenameInput) throw new Error('filename text input not found');
await page.mouse.click(filenameInput.x, filenameInput.y);
// Small settle so the focus click lands before typing — no JS-observable "input
// focused" signal here (documented interaction wait).
await page.waitForTimeout(200); // eslint-disable-line -- documented interaction dwell: focus click commit; no JS-observable focus signal
await page.keyboard.type(pcbFilename);
await page.waitForTimeout(300); // eslint-disable-line -- documented interaction dwell: typed-text registration has no observable signal
await page.keyboard.press('Enter');
// ── Wait for the load to complete (no dialogs visible). The
// wxInfoBar that pcbnew shows for "older format" PCBs is not a
// wxDialog, so waitForBoardLoaded doesn't get blocked by it. ──
// If we ever need to dismiss post-load wxMessageDialogs (missing
// libs etc.), do it INSIDE waitForBoardLoaded so the dismiss
// side-effect lives with the polling loop — calling page.evaluate
// from the test driver hangs if a runtime error breaks the wasm
// event loop (the asyncify-era post-load clipboard error was the
// proven case). ─────────────────────────────────────────────────
// ── Wait for the load to complete (no dialogs visible). ───────
const result = await waitForBoardLoaded(page, testLogger, 60000);
console.log(`[TEST] ${demo.name} board-ready result: ${result}`);
// Take the screenshot immediately. Don't wait — KiCad's post-load
// clipboard-polling path can hit a wasm RuntimeError that occasionally
// closes the page entirely on Firefox; if we sleep first we sometimes
// lose the page before page.screenshot runs. The canvas is already
// fully painted by the time waitForBoardLoaded returns.
await page.screenshot({
path: shotPath(page, `load-pcb-${demo.name}.png`),
scale: 'css',
});
const allLines = [...testLogger.consoleLogs, ...testLogger.errors];
const rtreeDiag = allLines.filter((l) => l.includes('[RTREE-DIAG]'));
expect(
rtreeDiag,
`RTree Classify duplicate-index reappeared for ${demo.name}; this means the wasm-layer fix at shape_poly_set.cpp:1927 has regressed:\n${rtreeDiag.join('\n\n')}`,
).toEqual([]);
const aborts = allLines.filter((l) => l.includes('Aborted('));
expect(
aborts,
`WASM aborted during ${demo.name} load:\n${aborts.join('\n\n')}`,
).toEqual([]);
// ── Clean-console gate: NO wasm trap may surface anywhere in the
// load — not before, not after the board renders. (Historical: the
// once-tolerated post-load clipboard/unwind RuntimeErrors were
// fixed in the asyncify era — sync clipboard IsSupported in wx,
// "unwind" sentinel handling — see docs/features/asyncify-arbiter/.)
const wasmTrapSignatures = [
'index out of bounds',
'indirect call to null',
'uncaught exception: unwind',
'invalid state',
'is not a function',
];
const wasmTrapErrors = allLines.filter((l) =>
wasmTrapSignatures.some((sig) => l.toLowerCase().includes(sig)),
);
expect(
wasmTrapErrors,
`wasm trap surfaced during ${demo.name} load:\n${wasmTrapErrors.join('\n\n')}`,
).toEqual([]);
});
}
test.describe('Load real PCB via File → Open', () => {
// Two 187 MB wasm runtimes loaded in parallel saturate Firefox's memory
// and slow each test enough that the 180s per-test budget runs out before
// the post-load canvas-settle wait completes. Run serially.
test.describe.configure({ mode: 'serial' });
test.setTimeout(240000);
for (const demo of DEMOS) {
runLoadPcbTest(demo);
}
});