Mark
4ea1bbb7a8
Session-2 (continuation of d614462 INCONCLUSIVE) executed disambiguation
plan and converged on a definitive verdict. Three independent observations
ruled out ALL architectural-failure hypotheses:
Step A: race-tolerant offscreen target attach (committed separately;
enabled visibility into the offscreen recorder + remux pipeline).
Step B: pre-kill and post-kill segment-count probes via the existing
`__mokoshOffscreenQuery 'get-segment-count'` bridge op (no new
test-only symbols introduced; FORBIDDEN_HOOK_STRINGS inventory
unchanged at 12 entries). Observed segments.length transition:
POST-PRIME=0 → PRE-KILL=3 → POST-KILL=3
Segments structurally survive the SW kill (offscreen still responds
to bridge query post-kill). Hypothesis A (architectural RAM loss
across SW termination) REFUTED.
Step C: SPIKE_SKIP_SW_KILL=1 env-var mode skips worker.close(). The
resulting videoSize is IDENTICAL to the canonical run (8505 bytes).
Hypothesis C (CDP-induced offscreen collateral teardown) REFUTED.
Since SW was not killed, its console listener stayed connected,
exposing the full Remux pipeline output:
[SW:Remux] Segment ts=1: 0 frames, duration=0ms, trackInfo=320x180
[SW:Remux] Segment ts=2: 0 frames, duration=0ms, trackInfo=320x180
[SW:Remux] Segment ts=3: 0 frames, duration=0ms, trackInfo=320x180
[SW:Remux] Remux complete: 0 frames, total timeline=0ms, output=8505 bytes
Each segment Blob has a valid track header (PixelWidth/Height parsed
successfully) but ZERO VP9 frames. Hypothesis B (canvas-captureStream
throttling in headless idle) CONFIRMED.
VERDICT: REFUTED-architecture (canvas-captureStream issue).
The architecture (offscreen-RAM `segments: Blob[] = []`) works
correctly; the spike's test methodology is invalid. The
`installFakeDisplayMedia` synthetic stream (canvas.captureStream(30)
on a hidden -9999px-offset 320x180 canvas) cannot sustain frame
production during a 5-min headless idle window despite the
`setInterval(drawFrame, 33ms)` belt-and-suspenders mitigation. This
matches the documented Chromium throttling of MediaRecorder on
invisible-canvas sources (Chrome bug 653548; auto-throttled-screen-capture
design doc; sendrec.eu blog "Why Canvas Breaks Your Screen Recorder").
ROUTING RECOMMENDATION (out of scope for this debug session):
- Do NOT proceed with the IndexedDB persistence plan-fix proposed by
Plan 04-04 SUMMARY. The plan-fix would NOT close SC #1 because the
spike would STILL produce 8505 bytes after IDB lands — the failure
is in the test's fake stream, not in segment persistence.
- Open a new plan slot (likely Plan 04-08 or a Phase 5 plan) that
reframes SC #1 verification methodology. Options:
(a) real getDisplayMedia in non-headless Puppeteer with
--auto-select-desktop-capture-source;
(b) video-file-backed MediaStream source (HTMLVideoElement
playing a bundled WebM) — bypasses canvas-captureStream
throttling entirely;
(c) reduce SC #1 wall-clock idle threshold to a value short
enough that canvas-captureStream survives (e.g., 30s) AND
add a separate manual operator-empirical test for 5-min.
ROADMAP SC #1 status: REMAINS OPEN. The architecture is sound; the
empirical verification gate is broken. Plan 04-04 SUMMARY's
characterization ("spike FAILED → architectural plan-fix needed") is
TECHNICALLY CORRECT on the first clause but INCORRECT on the second —
the spike's failure mode is in test infrastructure, not in production
code.
Files in this commit:
- tests/uat/spike-a33-sw-persistence.ts: added probeSegmentCount
helper using existing __mokoshOffscreenQuery bridge op; 3
checkpoints (POST-PRIME / PRE-KILL / POST-KILL); SPIKE_SKIP_SW_KILL=1
env-var skips worker.close() for Step C disambiguation.
- .planning/debug/sw-offscreen-persistence-investigation-session-2.md:
NEW session-2 debug note documenting full evidence trail + verdict
derivation + routing recommendation.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Mokosh UAT harness (Plan 01-11)
Puppeteer-driven Node script that runs 14 assertions end-to-end against a real Chrome instance loaded with the Mokosh extension. Replaces Plan 01-09 Task 5's operator-empirical functional verification (the operator retains only step 1 — build — and step 14 — brand/design acceptance).
Quick start
npm run test:uat
This builds dist-test/ (the hook-enabled bundle) and runs the harness.
Exit 0 means all 14 assertions passed. Final line: UAT harness: 14/14 assertions passed.
Local-debug mode
HEADLESS=0 npm run test:uat
Opens a real Chrome window so you can watch the picker auto-accept, the badge transitions, the popup appear, etc.
Developer iteration tricks
# Skip the production build inside assertion 0 (uses existing dist/):
SKIP_PROD_REBUILD=1 npm run test:uat
# Run the harness against an existing dist-test/ (skip npm run build:test):
npx tsx tests/uat/harness.test.ts
Assertion catalog
| # | Title | Bug class | Hook used |
|---|---|---|---|
| 0 | Production bundle has no test-hook leaks | T-1-11-01 | filesystem grep |
| 1 | SW bootstrap → setIdleMode | — | sw.evaluate |
| 2 | Toolbar onClicked-idle → REC + popup | — | triggerExtensionAction |
| 3 | Offscreen displaySurface === monitor | D-15 | __mokoshTest.getCurrentStream |
| 4 | Toolbar onClicked-recording → popup, no new offscreen | — | targets count |
| 5 | SAVE_ARCHIVE → download fires | — | downloads polling |
| 6 | BUG B: simulateUserStop → badge OFF + no recovery notif | b9eeeeb |
dispatchEvent('ended') |
| 7 | RECORDING_ERROR codec-unsupported → ERR + recovery notif | — | sendMessage |
| 8 | BUG A: onStartup → mokosh-startup- notification creates | a881bf0 |
__mokoshTest.handlers.onStartup |
| 9 | Icon file sizes meet floors | Bug A precondition | sw.evaluate(fetch) |
| 10 | Manifest has notifications + 3 icons | Bug A precondition | chrome.runtime.getManifest |
| 11 | 35s recording → segments.length >= 3 | D-13 | __mokoshTest.getSegmentCount |
| 12 | ffprobe on extracted webm exits 0 | Plan 01-08 | jszip + execFile |
| 13 | Archive shape — video + meta.json version match | Plan 01-07 | jszip |
Failure isolation
Single browser, serial assertions, bail on first failure for setup- dependent assertions (assertion 0 abort means refusing to launch a potentially-leaky bundle). Per-assertion bail keeps the diagnostic output unambiguous — see RESEARCH §5 + Plan 01-11 open-question resolution 4.
On failure, the harness dumps the last 30 lines of SW console + last 30 lines of offscreen console (captured live during the run) to stderr BEFORE rethrowing — gives you contextual triage without needing to re- run with debug logging.
Known gotchas
Locale-specific picker auto-accept
The --auto-select-desktop-capture-source=Entire screen Chrome flag
auto-accepts the screen-share picker. The string "Entire screen" is
en_US-specific. If your Chrome is set to a non-English locale, the
picker option label will differ and the auto-accept will silently fail
(picker stays open; assertion 2 times out).
Fallback: switch your Chrome user-data-dir's locale to en_US for
harness runs, OR adjust the launch arg in tests/uat/lib/launch.ts to
match your locale's equivalent string.
dev-dep Chromium binary size
puppeteer pulls a ~150 MB Chromium binary at npm install time. CI
must accept this. Production npm install --omit=dev skips it cleanly.
Xvfb is NOT required
Per Plan 01-11 RESEARCH §3 empirical probes against Chrome 148, the
--headless=new mode handles screen capture without Xvfb on Linux CI
runners. If a future Chrome regresses this, Xvfb :99 & DISPLAY=:99 npm run test:uat is the fallback.
CI runner screen-capture concern
The 35s recording assertion (A11) captures whatever is on screen during that window. CI MUST run the harness in an isolated container with no concurrent workload — see T-1-11-02 in Plan 01-11's threat model.
Real Chrome download (assertion 5 → A12)
The harness configures per-page download behavior via CDP to a fresh
os.tmpdir()/mokosh-uat-downloads-* directory; downloads are NOT
written to your real ~/Downloads. The temp directory is deleted by OS
tmpdir GC.