Repository navigation
Conversation
…erwrites the saved position After reopening, the test rendered a bare MessagesTimeline whose onManualNavigation was a no-op, so live follow stayed on. Once the restore settled, the next scroll save stored the manual position as atEnd: true, and the final switch back landed at the end. Render the reopened timeline through the Chat wrapper and turn off live follow on manual navigation, as ChatView does.
github-actions
Bot
added
vouch:trusted
![macroscopeapp[bot]](https://avatars.githubusercontent.com/in/900172?s=60&v=4){kind=small}
ApprovabilityVerdict: Approved at Macroscope's review found this PR approvable — This is a narrowly scoped browser-test-only flake fix confined to an ignored path, with no production runtime or product-default changes. Its impact is limited to test execution. Notes:
You can add or adjust custom eligibility rules. Learn more. |
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info
📝 Walkthrough
Merge Risk: ⚪ Minimal · up to
|
| Check name | Status | Explanation |
|---|---|---|
| Title check |  |
The title is concise, specific, and accurately describes the test-only fix for intermittent timeline restoration failures caused by a late live-follow save. |
| Description check | |
The description includes all required sections. It clearly explains the intermittent failure, the test-only change, scope and approval context, and focused verification results. |
| Linked Issues check | |
Check skipped because no linked issues were found for this pull request. |
| Out of Scope Changes check | |
Check skipped because no linked issues were found for this pull request. |
✨ Finishing Touches
-
🧪 Generate unit tests (beta)
-
- Create a new PR
-
- Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts
Comment @coderabbitai help to get the list of available commands.
|
Note Grok responding on behalf of Julius. Thanks for tracking this down. #22557 just landed on main and deleted the "restores a manual reading position after reopening and switching threads" test this PR fixes, so there's nothing left for it to change. That also takes care of item 5 in #22324. Closing as superseded. |
2 participants
Problem
MessagesTimeline.browser.test.tsx› "MessagesTimeline scroll position › restores a manual reading position after reopening and switching threads" fails intermittently withexpected 600 to be less than or equal to 1at the final switch back to the saved thread (#22324, item 5; also reported in #21968). The 600 is the test's wheel distance.rememberTimelinePositionstoresatEnd: liveFollowEnabled || position.atEnd. After reopening, the test rendered a bareMessagesTimelinewithbaseProps(), which hasliveFollowEnabled: trueand a no-oponManualNavigation. The restore effect callsonManualNavigation()for a manual position, but nothing turned live follow off. Once the restore settled, the nexthandleScrollframe saved the same offset asatEnd: true, so the final switch back scrolled to the end. The test only failed when that frame ran before it switched threads, which is why it passed on rerun.ChatViewwiresonManualNavigationtosetTimelineLiveFollowEnabled(false), so the app itself is not affected.Change
The reopened timeline now renders through the test's existing
Chatwrapper. The wrapper takesrouteThreadKeyand passesonManualNavigation={() => setLiveFollowEnabled(false)}, matchingChatView. The component is unchanged.The fix proposed in #22324 (reusing the wrapper alone) isn't enough: the wrapper only turns follow off on wheel, and nothing wheels after the reopen.
Scope and approval
A test-only fix for one flaky test. It's item 5 of #22324, and @juliusmarminge invited picking up the remaining #22324 flakes. This PR fixes only that item.
Verification
{"atEnd":true,"scrollOffset":3174}with the expected offset of 3174. The wrapper withoutonManualNavigationfails the same way. With this change, it passes.vp test run --project browser src/components/chat/MessagesTimeline.browser.test.tsx: 77/77 pass. The target test passed 10/10 repeated runs.Investigated and written by Claude Opus 5.5 in Claude Code (T3 Code).