Repository navigation
fix(web): collapsing the working window no longer nudges the timeline - #30466
Merged
Conversation
…s drop "Until" (pingdotgg#29134) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…dotgg#29980) * fix(web): project location keeps its machine name when the path truncates
…edesign colors (pingdotgg#29207) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…it (pingdotgg#29984) * fix(web): leave the wake-up fold answer unchanged
…ingdotgg#29998) * fix(web): a running wake-up's streaming answer replaces the previous one below its fold
maria-rcks
force-pushed
the
maria/redesign/2-composer-timeline
branch
from
November 4, 2026 06:07
9652e22 to
e908088
Compare
macroscopeapp
Bot
dismissed
their stale review
November 4, 2026 08:01
Dismissing prior approval to re-evaluate 633a668
maria-rcks
force-pushed
the
maria/redesign/2-composer-timeline
branch
from
November 5, 2026 01:04
e908088 to
80cc977
Compare
…-window-collapse-jump
maria-rcks
merged commit Nov 5, 2026
bd30925
into
pingdotgg:maria/redesign/2-composer-timeline
16 of 17 checks passed
SunkenInTime
pushed a commit
to SunkenInTime/t3code
that referenced
this pull request
Nov 5, 2026
SunkenInTime
pushed a commit
to SunkenInTime/t3code
that referenced
this pull request
Nov 5, 2026
SunkenInTime
pushed a commit
to SunkenInTime/t3code
that referenced
this pull request
Nov 5, 2026
SunkenInTime
pushed a commit
to SunkenInTime/t3code
that referenced
this pull request
Nov 5, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Note
🤖 Claude Opus 5.5 on behalf of Oliver
Fix for the redesign stack, targets layer 2 (#28840).
When you're at the bottom of a thread with a running turn and collapse the working window, the whole timeline jumps up about 10px for a few frames, then drops back. It happens whenever a new or updated row lands during the 200ms collapse, which happens almost every time during a busy turn.
The cause: the timeline pauses its "follow the end" scrolling only while the history is open, plus two frames after a toggle. That smooth follow comes back during the collapse. A new row then starts a smooth scroll to an end that is still moving up as the window shrinks. The browser then clamps the scroll back.
Now the working window tells the timeline when its collapse is running, and the timeline holds off following until the collapse transition settles. It already waited for that moment to swap in the collapsed box. Following comes back as soon as the collapse ends. It also comes back if the window unmounts partway through. The bar stays put, the timeline stays pinned to the bottom, and the collapse and its top fade play as before.
A new case in
MessagesTimeline.browser.test.tsxslows the collapse, adds a row partway through, and checks that the bar never moves. Without the fix, the bar jumps from 865px to 46px.Before / after
Local test setup, Replay only: a busy turn from a temporary Replay scenario (a long first answer, then a new row about every 150ms), with the collapse slowed to 3s by an injected
transition-duration. The slowdown gives rows time to land mid-collapse, so the bump shows up every time. 1280×831, dark mode. Each video expands and collapses the window three times. Both builds show the "Scroll to end" pill during a slowed collapse and "Jump to latest" in the collapsed box after a reload. The branch chip sayslayer2in both because that is the checkout name on the test machine.Before (layer 2):
https://gh-file-drop-api-prod-galwoqjslzlnws6s.oliver-boorstein.workers.dev/f/b4db11eaeb562f4d/30466-before-slow-bump.mp4
After:
https://gh-file-drop-api-prod-galwoqjslzlnws6s.oliver-boorstein.workers.dev/f/805a12f6be5778f6/30466-after-slow.mp4
Made with Claude Opus 5.5 in Claude Code (T3 Code).
takeover verification
merged current main and the current redesign target with merge commits, preserving the collapse fix and current panel defaults. the pr remains four files, +61/-3. no force push.
blacksmith
vp run check:changed FETCH_HEADpassed against the current target: format, lint, knip, typecheck, 180 focused tests, and 3 compiler checks. two independent code reviews found no issues. chrome exercised open, collapse, and rapid reopen through replay; the bar stayed at y=150 px. the focused browser regression covers arrivals during collapse.verified with gpt-6.1-sol through the codex harness.
Merged-head verification
Blacksmith
check:changedpassed, including 1,521 tests. Two independent reviewers approved633a66802d. The existing takeover comment records the checks and runtime limits.