pcbjam/tests/kicad/occ-export.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

273 lines
14 KiB
TypeScript

import type { Page } from '@playwright/test';
import * as fs from 'fs';
import * as path from 'path';
import { test, expect } from './fixtures';
import { waitForEditorReady, stableShot, settledShot } from '../e2e/utils/element-tracker';
import { injectFromSubmodule } from './utils/fs-inject';
import { openBoardProgrammatically } from './utils/board-ready';
import { findWxButton, clickWxButton, openStepExportDialog, dismissReportDialog } from './utils/wx-dialogs';
/**
* STEP export through the occ_service worker (docs/features/occ-split/):
* pcbnew.wasm carries no OCC — DIALOG_EXPORT_STEP's browser branch runs
* EXPORTER_STEP, whose WASM shadow suspends via EM_ASYNC_JS and hands the
* job to `globalThis.occService` (worker with its own OCC-linked module).
*
* Asserted end to end:
* 1. occ_service.{js,wasm} is NOT fetched at boot or board load — only the
* export click triggers it (the lazy-load boundary).
* 2. The (unchanged) export dialog drives the whole chain: menu → dialog →
* Export button → worker → STEP bytes.
* 3. The result is a real STEP file (ISO-10303-21 magic, non-trivial size),
* captured by the provider stub where the app would download it.
*/
const KICAD_VERSION_DIR = '10.0';
const PROJECT_DIR_MEMFS = `/home/kicad/documents/kicad/${KICAD_VERSION_DIR}/projects`;
const DEMO = { name: 'pic_programmer', dir: 'pic_programmer', stem: 'pic_programmer' } as const;
async function loadBoard(page: Page, testLogger: { consoleLogs: string[]; errors: string[] }): Promise<void> {
const pcbFilename = `${DEMO.stem}.kicad_pcb`;
const proFilename = `${DEMO.stem}.kicad_pro`;
await injectFromSubmodule(page, `kicad/demos/${DEMO.dir}/${pcbFilename}`,
`${PROJECT_DIR_MEMFS}/${pcbFilename}`);
await injectFromSubmodule(page, `kicad/demos/${DEMO.dir}/${proFilename}`,
`${PROJECT_DIR_MEMFS}/${proFilename}`);
const result = await openBoardProgrammatically(
page,
`${PROJECT_DIR_MEMFS}/${pcbFilename}`,
DEMO.stem,
testLogger,
60000,
);
console.log(`[TEST] ${DEMO.name} board-ready result: ${result}`);
}
test.describe('OCC export via occ_service worker', () => {
test.describe.configure({ mode: 'serial' });
test.setTimeout(240000);
test('export dialog produces a valid STEP; occ_service fetches lazily', async ({ page, testLogger }) => {
// Track occ_service fetches from the very start — the lazy boundary is
// the core assertion.
const occFetches: string[] = [];
page.on('request', (r) => {
if (r.url().includes('occ_service')) occFetches.push(r.url());
});
await page.goto('/kicad/pcbnew.html');
await waitForEditorReady(page);
await loadBoard(page, testLogger);
// Board-loaded ≠ board-painted: once the export dialog's modal pump takes
// over, the GAL may never repaint behind it — whichever paint state the
// canvas had at menu-open time is what the dialog screenshots freeze.
// Settle the pixels first so occ-export-{dialog,done}.png always capture
// the painted board (this bistability flagged occ-export-done between two
// identical-code CI runs).
await settledShot(page.locator('#canvas'), expect);
expect(occFetches, 'occ_service must NOT be fetched before the export').toHaveLength(0);
// File → Export → STEP/GLB/… and the unchanged export dialog.
await openStepExportDialog(page);
await stableShot(page, 'occ-export-dialog.png');
expect(await clickWxButton(page, 'Export'), 'Export button click').toBe(true);
// The provider stub captures the bytes where the app would download.
await page.waitForFunction(
() => ((window as any).__occExports?.length ?? 0) > 0,
null, { timeout: 180000 });
const exports = await page.evaluate(() => (window as any).__occExports as Array<{
name: string; size: number; magic: string;
}>);
console.log(`[TEST] captured exports: ${JSON.stringify(exports)}`);
expect(exports).toHaveLength(1);
expect(exports[0].name, 'download name comes from the dialog')
.toMatch(/\.step$/i);
expect(exports[0].magic.startsWith('ISO-10303-21'), 'STEP magic').toBe(true);
expect(exports[0].size, 'non-trivial STEP body').toBeGreaterThan(10_000);
expect(occFetches.length, 'occ_service was fetched lazily by the export')
.toBeGreaterThan(0);
// Dismiss the "Export complete" report dialog: it opens on top of the
// export dialog's modal lease (1 → 2), which IS its observable signal.
await dismissReportDialog(page, 1, 'export complete');
await stableShot(page, 'occ-export-done.png');
});
test('worker decode fault settles concurrent native wait and the next export recovers', async ({ page, testLogger }) => {
await page.goto('/kicad/pcbnew.html');
await waitForEditorReady(page);
await loadBoard(page, testLogger);
// The exact open Promise and title/paint helper have completed. Keep a
// byte-stable baseline before opening a nested submenu.
await settledShot(page.locator('#canvas'), expect);
await openStepExportDialog(page);
const probeBoard = fs.readFileSync(
path.resolve(__dirname, '..', 'fixtures', 'demo', 'demo.kicad_pcb'),
'utf8',
);
// Start one direct service request before the dialog's native request.
// Both wait on the same lazy worker boot, then both post without a host
// mutex. The harness injects a messageerror only when both are in the
// generation's pending map, and the real worker is then terminated.
await page.evaluate((boardText: string) => {
const runtime = globalThis as any;
runtime.__occServiceTestHooks.messageErrorWhenPendingAtLeast(2);
runtime.__occParallelResult = null;
void runtime.occService.request({
kind: 'export',
board: new TextEncoder().encode(boardText),
jobJson: JSON.stringify({ format: 'step', export_components: false }),
fileName: 'parallel-probe.step',
}).then((res: unknown) => { runtime.__occParallelResult = res; });
}, probeBoard);
expect(await clickWxButton(page, 'Export'), 'first native Export button click').toBe(true);
await page.waitForFunction(
() => (globalThis as any).__occParallelResult !== null,
null,
{ timeout: 30000 },
);
const parallelResult = await page.evaluate(
() => (globalThis as any).__occParallelResult as { ok: boolean; report?: string },
);
expect(parallelResult.ok, 'the parallel request must settle on generation failure').toBe(false);
expect(parallelResult.report, 'the exact messageerror reason reaches the caller')
.toContain('message decode failed');
// The C++ Export() request was the second request in the same failed
// generation. Its exact wx wait must close and show the native failure
// dialog instead of leaving the export handler parked.
await page.waitForFunction(() => {
const registry = window.wxElementRegistry;
if (!registry) return false;
const dialogs = registry.findAll({ visible: true })
.filter((el) => /Dialog/.test(el.typeName ?? ''));
return dialogs.length >= 2;
}, null, { timeout: 30000 });
await expect.poll(
() => page.evaluate(() => {
const scheduler = (globalThis as any).__wxScheduler;
return scheduler?.pendingWaits?.('occ') ?? -1;
}),
{ message: 'the native OCC wait must be completed by fail-all', timeout: 10000 },
).toBe(0);
const failed = await page.evaluate(() => {
const runtime = globalThis as any;
const rendered = runtime.wxElementRegistry?.findAllRendered?.({}) ?? [];
return {
labels: rendered.map((el: any) => el.label ?? el.text ?? '').filter(Boolean),
service: runtime.__occServiceTestHooks.snapshot(),
};
});
expect(failed.service.maxPending,
'two requests must coexist in one generation; worker requests are not serialized')
.toBeGreaterThanOrEqual(2);
expect(failed.service.requestsPosted,
'every provider entry must reach the real worker transport')
.toBe(failed.service.requestsStarted);
expect(failed.service.workerGenerationsStarted,
'the two parallel requests must share one worker generation').toHaveLength(1);
expect(failed.service.pending, 'fail-all must drain the failed generation').toBe(0);
expect(failed.service.retiredGenerations, 'the shared generation must be retired')
.toContain(failed.service.workerGenerationsStarted[0]);
expect(failed.service.activeGeneration, 'the failed slot must be cleared').toBeNull();
expect(failed.service.armed, 'the one-shot fault must be consumed').toBe(false);
console.log(`[TEST-OCC] native fault dialog labels: ${JSON.stringify(failed.labels)}`);
// Resolve the parent action before dismissing the child, then reuse its
// exact DOM identity. This prevents a label/geometry re-query from
// turning the handback race into a click on some replacement control.
const retryExport = await findWxButton(page, 'Export');
expect(retryExport, 'the original parent Export button must remain registered').not.toBeNull();
expect(retryExport!.x, 'the retry target has stable geometry').toBeGreaterThan(0);
expect(retryExport!.y, 'the retry target has stable geometry').toBeGreaterThan(0);
expect(await clickWxButton(page, 'OK'), 'dismiss native export failure').toBe(true);
// page.mouse.click() completes when the browser has delivered the OK
// input, not when the nested native modal has unwound. The parent
// export dialog is intentionally non-interactive until that exact
// child lease closes. Wait for the scheduler's observable modal
// count to return from {export + failure} to {export} before clicking
// through to the parent.
await expect.poll(
() => page.evaluate(() => {
const scheduler = (globalThis as any).__wxScheduler;
return scheduler?.pendingWaits?.('modal') ?? -1;
}),
{ message: 'the failure child must retire before retrying its parent', timeout: 10000 },
).toBe(1);
// The export dialog remains open. Its next request must create a fresh
// generation and complete through the actual OCC module. Reuse the
// captured geometry so a re-query can't land on a replacement control.
if (!retryExport) throw new Error('parent Export button disappeared before retry');
await page.mouse.click(retryExport.x, retryExport.y);
await expect.poll(
() => page.evaluate(() => (globalThis as any)
.__occServiceTestHooks.snapshot().requestsStarted),
{
message: 'the exact parent retry must enter the OCC provider once',
timeout: 10000,
},
).toBe(failed.service.requestsStarted + 1);
await expect.poll(
() => page.evaluate(() => (globalThis as any)
.__occServiceTestHooks.snapshot().activeGeneration),
{ message: 'the retry must boot a replacement worker generation', timeout: 30000 },
).toBe(2);
await page.waitForFunction(
() => ((window as any).__occExports?.length ?? 0) === 1,
null,
{ timeout: 180000 },
);
const recovered = await page.evaluate(() => {
const runtime = globalThis as any;
return {
exports: runtime.__occExports,
service: runtime.__occServiceTestHooks.snapshot(),
schedulerDead: runtime.__wxScheduler?.dead === true,
occWaits: runtime.__wxScheduler?.pendingWaits?.('occ') ?? -1,
};
});
expect(recovered.exports).toHaveLength(1);
expect(recovered.exports[0].magic.startsWith('ISO-10303-21'), 'retry returns real STEP bytes')
.toBe(true);
expect(recovered.exports[0].size, 'retry returns a non-trivial STEP file')
.toBeGreaterThan(10_000);
expect(recovered.service.activeGeneration, 'retry must own a replacement generation').toBe(2);
expect(recovered.service.requestsStarted,
'the parent retry must add exactly one provider entry')
.toBe(failed.service.requestsStarted + 1);
expect(recovered.service.requestsPosted,
'the parent retry must post exactly once to the replacement worker')
.toBe(failed.service.requestsPosted + 1);
expect(recovered.service.workerGenerationsStarted,
'the retry must create exactly one replacement generation')
.toHaveLength(failed.service.workerGenerationsStarted.length + 1);
expect(recovered.service.workerGenerationsStarted,
'the replacement generation must be the active one').toContain(2);
expect(recovered.service.pending, 'replacement generation must quiesce').toBe(0);
expect(recovered.schedulerDead, 'the worker failure must not terminalize the editor').toBe(false);
expect(recovered.occWaits, 'the replacement native OCC wait must quiesce').toBe(0);
// Dismiss the retry's "Export complete" report dialog via its modal
// lease (export dialog holds 1; the report raises it to 2).
await dismissReportDialog(page, 1, 'retry export complete');
});
});