test(browser): wait on the signal each test asserts in local-agent-activity and message-navigation by matt2e · Pull Request #3444 · block/buzz-app

matt2e · 2026-11-13T02:49:31Z

Why

Two journeys fail intermittently in CI because each waits on a signal that does not match what the test then asserts.

Test	Sightings
`local-agent-activity.spec.mjs`, "opens an active message's empty thread without losing its activity preview"	Browser journeys: 10 Nov, main, WebKit 18/18, 10 Nov, Chromium 9/9, 11 Nov, WebKit 18/18, 11 Nov, WebKit 18/18
`local-agent-activity.spec.mjs`, "finishing activity retires keyboard controls without stealing focus: stop"	Browser journeys: 10 Nov, WebKit 18/18
`message-navigation.spec.mjs`, "exact reveal uses ordinary dwell rather than marking read at open"	Browser journeys: 10 Nov, main, Chromium 5/9

Root cause and fix

One commit per spec.

Thread rows still positioning. While a thread positions its first page, ThreadPanel renders its rows with inert and sets aria-busy on the "Thread messages" region. Playwright's toBeVisible() ignores inert, and locator.press() and focus() make no actionability check. The two steps added in #3140 could therefore run before positioning finished.

In the empty-thread test, activity.press("Enter") went to the still-focused reply composer, which is prefilled with a mention. WebKit sent the draft and raised the "Mention people outside this channel?" dialog. Chromium kept focus in the composer. The Activity tab never appeared. The composer's toBeFocused() passes early because the composer sits outside the inert rows.
In the "stop" variant, focus() on Stop was a no-op, so the handoff to "More message actions" never happened.

DOM snapshots from the four traces examined show aria-busy="true" and inert on the rows at the moment of the press or focus. Every sighting post-dates #3140. This is not a product bug: people cannot reach inert rows either.

The tests now wait for the region's aria-busy="false", the readiness signal channel-lifecycle, nested-replies and navigation-thread-history already use. The wait sits after the composer toBeFocused() and before the Enter press, and in the finishing loop after the bubble is visible and before any focus(), so the bubble variant is covered as well.

Read publication under a default timeout. The exact-reveal dwell test ended by polling the fixture's read publications with the default 10 s timeout. A publication needs the 300 ms reading dwell, then read-state's 5 s publish debounce, then the flush's read, sign and publish. That leaves about 4.5 s of slack. In the failing trace the flush's read reached the fixture on time and the relay answered in about 4 ms, but the browser took 5.6 s to receive the response body. The signing request arrived 0.9 s after the poll deadline. The evidence file at teardown held the publication, so the behaviour was correct. The test was waiting on the wrong signal for what it asserts, which is that dwell marks read and open does not.

The test now asserts the local read with readJournal from reading.mjs, as the sibling test in the same file does. Right after the held query is seen, the journal's frontiers are still empty. At the end it polls for a non-empty frontier set. The journal write is strictly earlier than the publication and does not depend on the relay round-trip. The existing check that publications are unchanged at open is kept. The publication path stays covered by unread.spec.mjs, which polls with a 12 s timeout.

No timeouts or tolerances changed. No browser cases were added or removed.

Evidence

Local runs on Linux, bin/pnpm test:browser:ci --no-deps --reporter=list.

Scope	Result
`local-agent-activity.spec.mjs -g "opens an active message\|without stealing focus"`, WebKit, `--workers=1 --repeat-each=8` (4 tests × 8)	32/32 passed
Same, Chromium	32/32 passed
`message-navigation.spec.mjs -g "exact reveal uses ordinary dwell"`, Chromium, `--workers=1 --repeat-each=8`	8/8 passed
Same, WebKit, `--repeat-each=4`	4/4 passed
Whole `message-navigation.spec.mjs`, `--workers=2`	Chromium 22/22, WebKit 22/22

During triage, a test-only route that delayed the relay query by 1.5 to 2.5 s reproduced the inert-row failure deterministically: without the wait WebKit and Chromium each failed 2/2 with the CI signature, and with the wait all passed.

Not reproduced, not run

Neither failure reproduced on an idle host without that injected delay. The slow response body in the read-publication failure was pinned to CI load. The repeat runs guard against regression rather than reproducing CI timing.
No human has tested this yet, so buzz-review-completed is not set.

Left alone

The same default-timeout publication poll remains in "settings blocks reading and focus in a retained thread" in this file and in notifications.spec.mjs.
Open PRs fix(activity): keep live indicators to typing dots #3183, Show live agent work only at the end of its conversation #3001 and Keep replies inline with immediate previews and preserved drafts #2610 edit other tests in local-agent-activity.spec.mjs. Open PRs perf(threads): virtualize the thread panel's replies #2911, fix(messages): don't let a message reveal steal focus the person moved #2563, fix(channels): open the unread reply from sidebar Activity #2555, feat(unread): restore the Inbox's evidence from a device record on launch #2407 and fix(unread): keep the viewer's threads in the Inbox under busier channels #2401 edit other parts of message-navigation.spec.mjs. A trial merge of each against this change is clean.

🤖 Generated with Claude Code

…ing or focusing them While a thread positions its first page, ThreadPanel renders its rows with inert={positioning} and sets aria-busy={positioning} on the "Thread messages" region. Playwright's toBeVisible() ignores inert, and locator.press()/focus() make no actionability check, so the two steps #3140 added could run before positioning finished. In the empty-thread test, activity.press("Enter") went to the still-focused reply composer, which is prefilled with "@Local tester": WebKit sent the draft and raised the "Mention people outside this channel?" dialog, Chromium just kept focus in the composer, and the Activity tab never appeared. The toBeFocused() on the composer passes early because the composer sits outside the inert rows. The "finishing activity … stop" variant failed the same way: focus() on Stop was a no-op, so the handoff to "More message actions" never happened. DOM snapshots from all four failing traces show aria-busy="true" and inert="" on the rows at the moment of the press or focus. Not a product bug: users cannot reach inert rows either. Wait for the region's aria-busy="false" — the readiness signal channel-lifecycle, nested-replies and navigation-thread-history already use — after the composer toBeFocused() and before activity.press("Enter"), and in the finishing loop after the bubble is visible and before any focus(), so the bubble variant is covered as well. Verified with `bin/pnpm test:browser:ci --no-deps --workers=1 --repeat-each=8 --reporter=list tests/browser/local-agent-activity.spec.mjs -g "opens an active message|without stealing focus"` (4 tests × 8): - --project webkit: 32/32 passed (2.5 m; 4.1–4.2 s and 3.3–3.4 s). - --project chromium: 32/32 passed (2.1 m; 3.1 s and 2.5–2.7 s). The triage only reproduced the inert-row press by delaying the relay query, not on an idle host, so these runs guard against regression rather than reproduce CI timing. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: mtoohey <mtoohey@squareup.com> Signed-off-by: Matt Toohey <contact@matttoohey.com>

…er an exact reveal The exact-reveal dwell test ended by polling app.report.readPublications with the default 10 s expect timeout. A publication needs the 300 ms reading dwell (use-reading.ts), then read-state's 5,000 ms publish debounce (read-state.ts schedule()), then flush()'s own-slot read, sign and publish, leaving about 4.5 s of slack. In the failing CI trace the flush's read reached the fixture on time and the relay answered in ~4 ms, but the browser took 5.6 s to receive the response body, and read-state-sign arrived 0.9 s after the poll deadline. evidence.json at teardown held the publication, so the behaviour was correct; the test was waiting on the wrong signal for what it asserts (dwell marks read, open does not). Assert the local read instead, with readJournal from reading.mjs as the sibling test in this file already does: right after the held query is seen, the journal's frontiers are still {} (nothing is marked read at open), and at the end poll for a non-empty frontier set. save() awaits storage.update() before schedule() arms the debounce, so the journal write is strictly earlier than the publication and does not depend on the relay round-trip. The publication path itself stays covered by unread.spec.mjs, which polls readPublications with timeout: 12000. The existing readPublications-unchanged check at open is kept. Verified with `bin/pnpm test:browser:ci --no-deps --reporter=list tests/browser/message-navigation.spec.mjs`: - --project chromium --workers=1 --repeat-each=8 -g "exact reveal uses ordinary dwell": 8/8 passed (3.4–3.5 s each, 1.1 m). - --project webkit --workers=1 --repeat-each=4 -g "exact reveal uses ordinary dwell": 4/4 passed (4.3 s each, 36.8 s). - whole file, --workers=2: Chromium 22/22 (1.5 m), WebKit 22/22 (2.4 m). Biome is clean on the file. The slow body read did not reproduce on an idle host; the triage pinned it to CI load, so the local runs guard against regression rather than reproduce CI timing. Not addressed here: the same default-timeout readPublications poll after page.clock.resume() in "settings blocks reading and focus in a retained thread" (this file) and in notifications.spec.mjs. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Signed-off-by: mtoohey <mtoohey@squareup.com> Signed-off-by: Matt Toohey <contact@matttoohey.com>

kalvinnchau

🤖 One timing-dependent test assertion needs a controlled clock.

…ot race dwell Review of PR 3444 found that the empty-journal assertion added in eb5165c races the behaviour it sits next to. ThreadPanel's completeTarget sets readingSettled and dispatches reading-positioned before it completes navigation, so use-reading.ts starts the target's 300 ms dwell before openTarget() even resolves. The held depth query does not stop it: the exact target is already shown and focused. If the test needs more than 300 ms of real time to get from there to its IndexedDB read, correct behaviour has already written the target's frontier and `toEqual({})` fails. The check only passed because an idle host gets there in time. Confirmed before changing anything: with a temporary 400 ms wait after `await seen`, the eb5165c test failed 2/2 on Chromium and 2/2 on WebKit at that assertion, each time with exactly one frontier, `msg:<target id>: 1700000182`. Unmodified, it passed 2/2 on WebKit. Control the page clock instead, the way "settings blocks reading and focus in a retained thread" below already does: - Install the clock before the page loads, and pause it at a fixed offset once startup, unread.ensure() and the route are in place. - Start openTarget() without awaiting it and step 16 ms frames until navigation reports "opened". Exact reveal needs two animation frames, and Playwright's clock fakes requestAnimationFrame, so a paused clock would otherwise stall it. Dwell can have earned at most one frame. - Assert the journal is empty and readPublications is unchanged (both kept), and that the target is focused. Fake time is frozen here, so no amount of real time lets dwell elapse. - Advance 300 ms and poll the journal for exactly `{ "msg:<target id>": target.created_at }`. This replaces the old "some frontier exists" poll and is exact because the target is the only thread row while its context query is held. The poll only waits for the asynchronous IndexedDB write; it advances no time. - Release the held query, resume the clock, and keep the 81-row and focus checks. The held query, the route handler and navigation completion run on network and promise callbacks, which the clock does not fake. The sibling test's holdReadingFocus() is not needed: startup leaves focus on <main>, outside every reading surface, and the journal was `{}` at the pause in every run. No timeouts or sleeps were added, and nothing outside this test changed. Verified on a loaded local macOS host (load average 34 to 42) with `bin/pnpm test:browser:ci --no-deps --workers=1 --reporter=list tests/browser/message-navigation.spec.mjs -g "exact reveal uses ordinary dwell"`: - --project=chromium --repeat-each=8: 8/8 passed (1.2 m, 3.4 to 4.1 s). - --project=webkit --repeat-each=8: 8/8 passed (1.5 m, 3.8 to 6.1 s). - With the temporary 400 ms wait after `await seen` in the fixed test, --repeat-each=2: Chromium 2/2 and WebKit 2/2 passed. The wait was removed before committing. - A temporary dump of the fake timers just before the 300 ms advance (5 Chromium runs, 3 WebKit) showed the same state every time: clock at 20,064 ms, one thread row, dwell timer due at 20,364 ms, and after the advance the 5 s publish debounce armed at 20,364 ms. The dwell was never restarted after "opened". - Biome is clean on the file; `git diff --check` is clean. Not run: the rest of message-navigation.spec.mjs, any other spec, and Linux engines. Hosted CI covers those. Signed-off-by: Matt Toohey <contact@matttoohey.com> Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>

kazemek21

Kyle reviewed the PR and the Nexus review-tracker notes and approved with no comments.

Posted by Codex (gpt-6.1-sol) on behalf of kylez.

wesbillman

🤖 Star Lord’s automated source review via Wes’s account. Reviewed 0ab64f7d313453b608fda90ac7d59a807d6dae5b against 7f03764beb8eb0ddcb4ff9ec0b8bda3e07699716.

One P2: the exact-reveal clock repair controls target dwell, but startup can still earn a channel read before the clock is paused. Details and the existing helper to reuse are inline. Thread readiness waits match the state controlling inert rows, and the target-frontier assertion preserves the intended local durability boundary.

Validation: source review only; the startup race was not reproduced in this review, and no tests or app execution were performed. Existing hosted run 43379266580 records 10 passing repetitions of each of the five changed scenarios in both Chromium and WebKit (100/100). Artifacts bind to clean merge 641d207adf54abf0ced682bccbb2b961a221578a, containing this head and earlier target 7c786bb464a72fd7ea25084a9f57edd4b347093d; these passes do not establish the disputed startup ordering or validate the current-base combination. No browser cases were added or removed. Timing artifacts were inspected; no matched before/after baseline was established.

wesbillman · 2026-11-13T12:25:15Z

+    // only a paused clock keeps "nothing read at open" from racing that dwell.
+    // Pause at a fixed offset, not the page's already-ticking clock.
+    const base = Date.now();
+    await page.clock.install({ time: base });


🤖 [P2] Hold startup reading focus before opening the app

page.clock.install() leaves time running until pauseAt below. During open(page, app), Messages initially autofocuses the channel composer, which qualifies for the normal 300 ms reading dwell. If a slow runner leaves that focus in place before the search palette takes over, Alpha can already save a frontier. Pausing afterward cannot undo it, so the empty-journal check and later target-only equality fail despite correct behavior. unread.spec.mjs:130–133 documents and guards this same startup ordering. Reuse holdReadingFocus(page) before open() and releaseReadingFocus(page) after pauseAt, as the neighboring retained-thread test does. This is a source-derived ordering defect; it was not reproduced in this review.

…y-unreads * origin/main: (24 commits) [1/2] feat(agents): add a recommended agent setup (core) (#3344) test(browser): survive the live-stream restart a page-clock jump causes (#3442) chore(renovate): hold rusqlite below 0.41 for mesh-llm (#3438) test(browser): serve the diffs and CLI reload fixtures outside the navigation budget (#3445) test(browser): play the draft clip at rate 0 instead of looping it before the Play click (#3446) fix(status): reopen a status whose preset deadlines coincide as Today (#3449) fix(messages): tolerate a cancelled reaction-name chunk load (#3450) fix(pairing): tolerate /proc/net/tcp rows printed twice or missed (#3451) test(browser): wait on the signal each test asserts in local-agent-activity and message-navigation (#3444) Expose native channel headings to Solo sources (#3523) test(browser-host): run the 100,000-channel broker scale check alone in the measurements lane (#3447) fix(solo): keep every contributed destination action (#3516) fix(identity): report an issuer that keeps rate-limiting an agent as a limit (#2365) fix(identity): let a stuck at-launch sign-in copy its link (#2453) fix(profiles): open avatar previews on hover in compact density (#3510) feat(solo): smooth chat recency reordering (#3502) Reject self-mentions that cannot wake another session (#3513) feat(status): offer don't clear for every status (#3515) Let plugin panels name their tab with panels.useTabTitle (#3251) fix(agents): report a conversation wake that fails before its turn starts (#2192) ... # Conflicts: # tests/browser/message-management.spec.mjs

* origin/main: (27 commits) [2/2] feat(agents): offer Automatic (Recommended) in Create agent and Identity (#3345) fix(solo): preserve and resume unsent chat drafts (#3525) fix(builderlab): give refused app sign-ins and Identity links a route to Organization sign-in (#3078) fix(agents): say why Import an agent can't import here, and link Manage plugins (#3073) [1/2] feat(agents): add a recommended agent setup (core) (#3344) test(browser): survive the live-stream restart a page-clock jump causes (#3442) chore(renovate): hold rusqlite below 0.41 for mesh-llm (#3438) test(browser): serve the diffs and CLI reload fixtures outside the navigation budget (#3445) test(browser): play the draft clip at rate 0 instead of looping it before the Play click (#3446) fix(status): reopen a status whose preset deadlines coincide as Today (#3449) fix(messages): tolerate a cancelled reaction-name chunk load (#3450) fix(pairing): tolerate /proc/net/tcp rows printed twice or missed (#3451) test(browser): wait on the signal each test asserts in local-agent-activity and message-navigation (#3444) Expose native channel headings to Solo sources (#3523) test(browser-host): run the 100,000-channel broker scale check alone in the measurements lane (#3447) fix(solo): keep every contributed destination action (#3516) fix(identity): report an issuer that keeps rate-limiting an agent as a limit (#2365) fix(identity): let a stuck at-launch sign-in copy its link (#2453) fix(profiles): open avatar previews on hover in compact density (#3510) feat(solo): smooth chat recency reordering (#3502) ... # Conflicts: # src-tauri/src/federated_identity/agent/tests.rs # src-tauri/src/federated_identity/mod.rs # src-tauri/src/federated_identity/protocol.rs

* origin/main: (133 commits) Expose ordinary chats without workspace headings (block#3530) [2/2] feat(agents): offer Automatic (Recommended) in Create agent and Identity (block#3345) fix(solo): preserve and resume unsent chat drafts (block#3525) fix(builderlab): give refused app sign-ins and Identity links a route to Organization sign-in (block#3078) fix(agents): say why Import an agent can't import here, and link Manage plugins (block#3073) [1/2] feat(agents): add a recommended agent setup (core) (block#3344) test(browser): survive the live-stream restart a page-clock jump causes (block#3442) chore(renovate): hold rusqlite below 0.41 for mesh-llm (block#3438) test(browser): serve the diffs and CLI reload fixtures outside the navigation budget (block#3445) test(browser): play the draft clip at rate 0 instead of looping it before the Play click (block#3446) fix(status): reopen a status whose preset deadlines coincide as Today (block#3449) fix(messages): tolerate a cancelled reaction-name chunk load (block#3450) fix(pairing): tolerate /proc/net/tcp rows printed twice or missed (block#3451) test(browser): wait on the signal each test asserts in local-agent-activity and message-navigation (block#3444) Expose native channel headings to Solo sources (block#3523) test(browser-host): run the 100,000-channel broker scale check alone in the measurements lane (block#3447) fix(solo): keep every contributed destination action (block#3516) fix(identity): report an issuer that keeps rate-limiting an agent as a limit (block#2365) fix(identity): let a stuck at-launch sign-in copy its link (block#2453) fix(profiles): open avatar previews on hover in compact density (block#3510) ... Signed-off-by: Joah Gerstenberg <joah@squareup.com> # Conflicts: # src/bundled/codex/prompt.test.ts # src/bundled/goose/runtime.ts

matt2e and others added 2 commits November 13, 2026 13:39

matt2e requested a review from a team as a code owner November 13, 2026 02:49

kalvinnchau reviewed Nov 13, 2026

View reviewed changes

Comment thread tests/browser/message-navigation.spec.mjs Resolved

kazemek21 approved these changes Nov 13, 2026

View reviewed changes

wesbillman reviewed Nov 13, 2026

View reviewed changes

matt2e merged commit 6896e9e into main Nov 14, 2026
55 checks passed

matt2e deleted the mtoohey/flake-readiness-signals branch November 14, 2026 00:39

Navigation Menu

Repository navigation

test(browser): wait on the signal each test asserts in local-agent-activity and message-navigation - #3444

test(browser): wait on the signal each test asserts in local-agent-activity and message-navigation#3444
matt2e merged 3 commits into
mainfrom
mtoohey/flake-readiness-signals

matt2e commented Nov 13, 2026

Uh oh!

kalvinnchau left a comment

Uh oh!

Uh oh!

kazemek21 left a comment

Uh oh!

wesbillman left a comment

Uh oh!

wesbillman Nov 13, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

matt2e commented Nov 13, 2026

Why

Root cause and fix

Evidence

Not reproduced, not run

Left alone

Uh oh!

kalvinnchau left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

kazemek21 left a comment

Choose a reason for hiding this comment

Uh oh!

wesbillman left a comment

Choose a reason for hiding this comment

Uh oh!

wesbillman Nov 13, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants