Skip to content

perf(chat): stop re-rendering the page on every keystroke and pin - #14227

Merged
bolichen97 merged 1 commit into
kirodotdev:mainfrom
Pearcekieser:perf/page-wide-redraws
Sep 29, 2026
Merged

bolichen97 merged 1 commit into
kirodotdev:mainfrom
Pearcekieser:perf/page-wide-redraws

Conversation

@Pearcekieser

@Pearcekieser Pearcekieser commented Sep 26, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Typing in the chat composer and pinning a session no longer re-render the whole page, so keystrokes land without visible lag.

Typing in the main chat re-renders most of the page on every key. A React commit probe counted 763 components rendered per keystroke on a 200-message session (3.4 commits per key), including the action menus of every visible message row. At 4x CPU throttle a keystroke took 75 ms to reach the next frame (p50, 176 ms p95), and 7 of the ~11 seconds spent typing 120 characters were long tasks. Pinning a session rendered 11,076 components, most of them sidebar row menus.

Why it matters

The composer is the surface people use most, and on slower machines or a busy page (a streaming reply, split panes) this is visible input lag. T3 Code, measured on the same host with the same harness, lands a keystroke in 17-24 ms at 4x.

Not a goal

  • Per-session WebSocket sidebar deltas and the server-side session-list push; another PR owns those.
  • Bundle size and first-load time; another PR owns those.
  • Narrowing the ~23 whole-session-list store subscriptions; they are under 2% of a pin's rendered components.
  • The per-keystroke costs left inside the composer (the automation popover's useIsFetching, inline dashboardConfig query functions, i18nT lookup caching); a stacked follow-up PR covers them.

What changed (motivation → approach → change)

Before After
Composer text is useState in ChatPage, so each key re-renders the page Text lives in a small store; only the composer subscribes
ChatInput gets ~20 fresh inline callbacks and JSX props per page render, so its memo never holds Callbacks go through stable forwarders, object props are memoized
A displaced sidebar row rebuilds its ⋯ dropdown and context menu Row menus are memoized; the row still re-renders for its layout spring

The draft moves into chat-core/composer/draftStore.ts, the P3-f step the Composer RFC plans. <Composer draft={store}> hands the store to ChatInput (components/ChatInput.tsx), which reads the text with useComposerDraftText(). ChatPage (pages/ChatPage.tsx) keeps its setInput call shape and reads the text through inputRef when it acts. Where it renders from the text, it selects one fact (blank or not, staged @dir/ tokens), so it re-renders only when that fact changes; pages/chat/useChatPageSessionController.ts takes the blank-or-not flag instead of the text. The Slack-link prompt hand-off writes the composer directly only once the composer is on the target session; mid-switch, the slot-change effect restores the prompt from the prefill it already stores, so the prompt never lands in the session being left. A child component, ComposerDraftSync, mirrors the committed text into inputRef and persists the slot draft from its effect. Child effects run before the page's own, so a keystroke batched with a slot switch still lands in the slot it was typed in.

StableChatInput wraps ChatInput with useStableCallbackProps (page-local, pages/chat/useStableCallbackProps.ts, listed in docs/feature-map/README.md as an internal helper), which swaps each function prop for a forwarder that calls the latest one. The composer's aboveComposer content, session-control chips and refused-options set are memoized. hooks/useSessionControls.ts stops handing out a fresh [] and a fresh statuses map on every render, since those broke the memo chain. In pages/ChatSidebar.tsx, SessionRow memoizes its menu elements, re-keyed on a language change, because a row re-renders whenever its paint position moves.

The ChatInput message-list subscription from the original plan is already gone on main (it reads messages at click time), so nothing changed there. Replacing every whole-slot-list selector was measured and left out: ChatPage's own share of a pin is under 2% of rendered fibers. What remains of a pin is framer-motion's layout spring on the displaced rows plus the full slot-list push, which the sidebar-deltas work covers.

Backwards compatibility

Compatible: ChatInput's value becomes optional and is still honoured without a draft root; Composer accepts value as before. No persisted format changes.

Tests

  • draftStore.test.tsx: setter shape and no-op writes; a selector caller re-renders only when its fact flips.
  • Composer.test.tsx: under <Composer draft>, ChatInput shows and edits the store text while the host renders zero times.
  • useStableCallbackProps.test.tsx: a memoized child skips a host render with fresh arrows, calls the latest closure, and keeps an absent handler absent.
  • ChatSidebar.rowMemo.test.tsx: after a top insertion every displaced row re-renders but only the new row builds a context menu (42 before this change, 1 after).
  • ChatPageDrafts.test.tsx: the effect-order guard follows the text persist into ComposerDraftSync.
  • ChatPage.persist.integration and ChatPage.sendDuringDictation: stand-ins read the text the way the real composer does.

Manual verification

Measured with the benchmark harness from the KiroCrew vs T3 Code comparison: a production build, a test-mode gateway with 67 seeded sessions, headless Chromium, three runs, medians. Same host and seed before and after.

typing, s200 (200 messages), 120 keys        before    after
components rendered per key (React probe)       763      161
commits per key                                3.37     1.35
main-thread busy ms per key (1x CPU profile)   26.3     11.0
key to next frame, 4x CPU, p50 / p95 ms     75 / 176  26 / 52
long tasks while typing, 4x, ms               7,030      420
key to next frame, 1x, p50 ms                  17.0     11.6

pin a session (right-click, Pin), 10 reps    before    after
components rendered per pin                  11,076    4,314
pin icon visible, 1x ms                        71.5     54.3
long tasks, 4x ms                               843      720

s3000 and the empty landing chat show the same typing shape (s3000 4x p50 80 to 27 ms). 4x p50 swings between reps, so the component and busy-ms rows are the steadier comparison.

Screenshots / video

Why no screenshot: render-scheduling change only; the composer, transcript and sidebar draw the same pixels.

Related Issues

no linked issue: follow-up from the KiroCrew vs T3 Code responsiveness comparison.

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

@Pearcekieser
Pearcekieser requested review from a team and cixuuz September 26, 2026 13:28
@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Intent: Keep a keystroke in the main chat composer, and a sidebar pin, from re-rendering the rest of the page, so typing stays responsive on slow CPUs and while replies stream.
Not a goal: Server-side slot-list pushes or metadata persistence, first-load bundle size, or changing the sidebar's row-slide animation.

@github-actions github-actions Bot added fork Pull request from a fork (external contributor) readiness: action required A blocking check or review needs attention readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Sep 26, 2026
@github-actions

github-actions Bot commented Sep 26, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Fable 5, fork) — ✅ PASS

UX-level review of d870e8ff24d6196596e2fd1d5440b47f09d38c5f via the fork AI-review pipeline — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

UX-Verdict: PASS

Verify a draft typed during a session switch (and the Slack-link prompt hand-off) still lands in the intended session, not the one being left.

[UX-REVIEWED] d870e8f

Description read

[DESCRIPTION-READ] a1807352e9e6dce693ed6d8029978bcdf70493ad6671c3484a1393b39c1b2579

That sha256 names everything this verdict read: the PR description, plus the 3 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Sep 26, 2026 •

Copy link
Copy Markdown
Contributor

GPT 5.6 Review (fork) — ✅ no blocking findings

Reviewed d870e8ff24d6196596e2fd1d5440b47f09d38c5f via the fork AI-review pipeline; updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] d870e8f

@github-actions

github-actions Bot commented Sep 26, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Fable 5, fork) — ✅ PASS

Design-level review of d870e8ff24d6196596e2fd1d5440b47f09d38c5f via the fork AI-review pipeline — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

Design-Verdict: PASS

Verify the Slack-link prompt mid-switch hand-off by hand: its restore-from-prefill path is the one behavior change no listed test exercises.

[DESIGN-REVIEWED] d870e8f

Description read

[DESCRIPTION-READ] bdb3c03871bc00a732880c875d0fcfe0145508a84f6f4ec70d78015e64887806

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Sep 26, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Fable 5, fork) — ✅ PASS

Premise-level review of d870e8ff24d6196596e2fd1d5440b47f09d38c5f via the fork AI-review pipeline — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS/CONCERNS are advisory.

First-Principles-Verdict: PASS

Verify the Slack-link hand-off manually: mid-switch the prompt now relies entirely on the slot-change effect's prefill restore (ChatPage.tsx prefill guard), a timing path no added test pins.

What this change ships

Inventory (10 items) — 10 justified

Intent: make typing and pinning stop re-rendering the whole chat page — a FIX for a measured performance defect (763→161 components/key, benchmark reproduction in the description; render-count tests added).

  1. A keystroke re-renders only the composer, not the page (text moves to a draft store) — justified
  2. Pinning no longer rebuilds every displaced sidebar row's ⋯/context menus — justified
  3. Composer accepts a draft store; value becomes optional and still honoured (ChatPane, its other host, keeps it) — justified
  4. ChatPage's composer callbacks pass through stable forwarders so memo(ChatInput) holds — justified
  5. Session-controls hook returns stable identities for its empty list and statuses map — justified
  6. Slack-link prompt mid-switch now lands via the prefill restore instead of a direct write — justified
  7. History suggestions and file-chip reconciliation fire from draft commits, not page renders — justified
  8. Above-composer content, control chips and refused-options set memoized — justified
  9. Effect-order test pin rewritten to follow the persist into ComposerDraftSync, with the child-effect rationale — justified
  10. Feature-map row updated in the same commit (AGENTS.md mandate) — justified

Lens checks run: the new store is meaningfully different from Composer.tsx's module-private createSlot (updater-form set with 4 existing setInput(prev => …) call sites; no notify/silent), so not a second spelling. draft has 1 consumer, useComposerDraft* hooks 2–3 each, useStableCallbackProps 1 (the singular form — ~20 hand-written forwarders — is larger, so no premature-generalization subtraction exists). Deferred siblings (composer-internal per-key costs, sidebar deltas) are named in ## Not a goal. RFC rfc-chat-core-extraction is partial on base and plans P3-f; no default or capability changes, so no product-shape trigger.

[FIRST-PRINCIPLES-REVIEWED] d870e8f

Description read

[DESCRIPTION-READ] 2366fc4d4838285ea761f363871568a6b7e6a41dc1635d229aa7bcabc1dc7ee1

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Sep 26, 2026
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Sep 26, 2026
@github-actions

github-actions Bot commented Sep 26, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5 Review (fork) — ✅ no blocking findings

Reviewed d870e8ff24d6196596e2fd1d5440b47f09d38c5f via the fork AI-review pipeline; updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] d870e8f

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention and removed readiness: checking Automated validation is still running labels Sep 26, 2026
@Pearcekieser

Copy link
Copy Markdown
Contributor Author
  • rebutted: keep the forwarders at the ChatPage call site

Only ChatPage builds ~20 inline callbacks per render; ChatPane, embeds and tests pass stable handlers, so moving the forwarders into ChatInput changes every surface to fix one.
useStableCallbackProps documents the contract where it is applied: the host's latest closure is always called, so an effect inside ChatInput keyed on a callback identity sees a stable forwarder that still runs current code.
No ChatInput effect relies on a callback identity changing to re-run; the three callback-keyed hooks are useCallbacks that forward.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author
  • rebutted: createSlot is kept private on purpose, and the two stores differ

createSlot exists for render-time writes: its set(next, silent) swaps without notifying and notify() is called separately from a layout effect; the draft store must never write silently, since a silent write would leave the editor showing stale text.
The draft store is a public seam (ChatPage, ChatInput, tests) with the useState-setter updater shape, plus the selector hook; exporting createSlot would put the silent-write path on that public surface.
Wrapping one in the other saves ~10 lines and couples the Voice atom's render-phase protocol to the editor text store.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author
  • fixed: the Slack-link prompt step is now in the description

The PR body's What changed section names the step: the token flow writes the composer only when the composer is already on the target session, and otherwise leaves it to the slot-change effect, which restores the prompt from the prefill hand-off the flow already writes.
Why it is needed: with the draft in a store, a write that commits in the same batch as the switch would persist the prompt into the session being left (ChatPageCoverage "keeps text typed during the token create in the old session" pins it).

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: main fixed the inherited settings-registry reds (#14138, f6d8b7f -> 7b2da39); ChatPanel.railRegistry and settingsRegistry pass locally on the new head ec65b13. No code change; CI re-running.

@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention labels Sep 26, 2026
@Pearcekieser

Copy link
Copy Markdown
Contributor Author
  • rebutted: the boundary is pinned where it can be pinned without a full-page harness

useStableCallbackProps.test.tsx pins the mechanism: a memoized child does not re-render on fresh inline callbacks, still calls the latest closure, and keeps an absent handler absent.
A ChatPage-level render-count pin needs the whole page mounted with a counting ChatInput stand-in; the ChatPage suites already mock ChatInput, so a counter there would count renders of a stub fed by the real page props, one more copy of the ChatPage harness for a guard the React profiler numbers in the PR body already show.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

self-added: yes
mechanism: followUpRefusedOptions useMemo (added in round 0 to keep ChatInput's memo)

  • fixed: the refused-options memo re-keys on the published latch set

Legitimate: isRefused is useCallback([slot]) and reads module-level latches, so the memo keyed on its identity missed the publishLatches update after a Go / Cancel click and Go All stayed undimmed.
Fix in eb80340: the memo depends on planActionMutation.latchedActions (the useSyncExternalStore snapshot that changes exactly when a latch publishes), so the Set is rebuilt on the same render the page already takes for that publish.

@github-actions github-actions Bot removed the readiness: checking Automated validation is still running label Sep 27, 2026
@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: main conflicted with this PR in ChatSidebar.tsx and docs/feature-map/README.md after #14462. Resolved with no behaviour change: the two memoized row menus now gate on main's new foreignRow flag instead of peerId, and the feature-map row keeps main's text plus this PR's useStableCallbackProps.ts entry. tsc, eslint, the feature-map gate and 102 related frontend test files (1,161 tests) pass on b59e528.

@iamwhatever the rebase dropped your approval and auto-merge. Could you re-approve and re-enable auto-merge once CI is green? Thanks.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

CI status on b59e528. Nothing here needs a code change in this PR:

  • PR Hygiene and Fork PR Description failed because main's template (feat(prepare-pr): freeze a PR's goal sections at open #14181) now requires a Goal: line and a ## Not a goal section. I added both to the description; .github/scripts/pr-description-check.sh passes on it.
  • Bundle Size Gate: chunk t is 28 B over its 955 KB budget. A local build of this head and of its parent commit produce a byte-identical t chunk (same hash, 970,797 B), so this diff adds nothing to it; the growth is on main.
  • Electron Shell Tests (no probe CALL is unclaimed by a list, no sentinel literal is unclaimed by a list) and Frontend Tests (2) (bnStyle.test.ts) fail the same way on main's CI run 36337451146.
  • Frontend Tests (4) died on a vitest worker crash ("Worker exited unexpectedly"), and Backend Tests (3.12, 5) failed in test_wake_judge_feedback.py, which this frontend-only PR does not touch.

I'm not pushing again until main's reds are fixed, since a new head needs fresh run approvals and resets review. A maintainer rerun of Frontend Tests (4) and Backend Tests (3.12, 5) would help once main is green.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: onto main f57c1de, whose CI is now green (the Bundle Size, Electron Shell and bnStyle reds on the previous head came from main). The rebase applied cleanly and changes no code. tsc, eslint on the changed files, and 484 related frontend test files (7,873 tests) pass locally on 9153bdb.

@iamwhatever, once CI is green could you re-approve and re-enable auto-merge? The rebase dropped both. The fork's workflow runs may also need approval first.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Re-trigger with no code change: e83af4f has the same tree as 9153bdb (dc7bb55220), with a new commit date. On 9153bdb every check passed except one. Backend Tests (Windows) (2) failed with a TimeoutError in test_auto_improvement_members.py::test_authors_retain_completed_edits_after_member_deadline[author_bug_fix]. This frontend-only PR doesn't touch that backend test, and PR Readiness was red only because of it.

@iamwhatever, once CI is green could you approve the workflow runs if needed, re-approve the PR and re-enable auto-merge?

iamwhatever
iamwhatever previously approved these changes Sep 28, 2026
@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: e83af4f had become CONFLICTING with main after the approval, so this is now 448dd7b on main b7ad967. The only conflict was in ChatPage.tsx, where main's #14472 added the mobile top-bar hooks next to this PR's langGen line. I kept both, with no behaviour change. None of the memoized composer or sidebar blocks read the new top-bar values.

Checked locally: tsc, eslint on the changed files, the feature-map gate, and 235 related frontend test files, all passing.

@iamwhatever, the push dismissed your approval again. Once the fork's workflow runs are approved and CI is green, could you re-approve and turn auto-merge back on? Sorry for the extra round.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

CI status on 448dd7b: three backend shards are red, and none of the failures comes from this frontend-only diff.

  • Backend Tests (3.12, 4) and (Windows) (4) fail test_mcp_tool_titles.py::test_every_listed_tool_carries_its_declared_title[kirocrew-dashboard]. It fails the same way on a clean checkout of main at dceaca3, so main has this failure too. It looks like feat(mcp): declare Tool.title on the own MCP servers' tools #14508, which declares a title on every tool, crossed with a dashboard tool that landed alongside it without a title.
  • Backend Tests (Windows) (6) fails test_fence_marker_separators.py::test_forwarded_fence_long_separator_run_is_linear on its 0.5 s wall-clock bound (0.55 s measured), which is a timing flake.

I'm not pushing while main has the failure. Once main has a fix, I'll rebase and push once, which re-runs every shard, including the flaky one.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: main fixed the test_mcp_tool_titles failure in #14641, so this branch is now 1efddff on main 631f8e6. The rebase applied cleanly, with no code change. Checked locally: tsc, eslint, the feature-map gate and 235 related frontend test files pass, and the title test passes on main. This run also re-runs the flaky Windows shard and the incomplete Opus review.

@iamwhatever, could you approve the fork's workflow runs, then re-approve and turn auto-merge back on once CI is green?

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

CI status on 1efddff: everything passes except two Windows backend shards. Both failed on timing assertions in backend tests that this frontend-only diff does not touch:

  • Backend Tests (Windows) (4): test_diag_procs.py::test_two_reads_too_close_together_decline_the_rate
  • Backend Tests (Windows) (6): test_fence_marker_separators.py::test_forwarded_fence_long_separator_run_is_linear (0.54 s against its 0.5 s bound)

Both shards passed on main's CI at the same base, 631f8e6. PR Readiness is red only because of these two.

I'm not pushing another same-tree commit, because it would need fresh run approvals and restart the review lanes. Could a maintainer use "Re-run failed jobs" on run 36389971127 for these two shards? @iamwhatever, once they pass, could you re-approve and turn auto-merge back on?

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Re-triggered CI with a same-tree push: e27ef12 has the same tree as 1efddff, with only the commit date changed. The only red checks on 1efddff were two Windows backend timing flakes, test_diag_procs and test_fence_marker_separators. This PR doesn't touch either test, and both shards pass on main at the same base. No rerun had happened after 6 hours, and a fork contributor can't rerun jobs.

@iamwhatever, could you approve the workflow runs, re-approve the PR once CI is green, and turn auto-merge back on? Thanks.

iamwhatever
iamwhatever previously approved these changes Sep 28, 2026
@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: onto main a25b321 (new head 552db76) to clear a conflict with #6511 (file-chip reconciliation) in ChatPage.tsx and the feature map.

#6511 added an effect keyed on the composer text ([input]). This PR moved that text into the draft store, so the page no longer has an input value. The reconciliation body is unchanged; it now runs from the draft commit (onComposerDraftCommit), once per committed text change, and reads the committed pendingFiles as the effect did. restoreQueuedDraft keeps main's mergeSlotTokens dependency and this PR's setInput.

tsc and eslint are clean, the feature-map gate passes, and 245 related frontend test files (3,072 tests) pass, including ChatPage.dirStaging, #6511's reconciliation suite.

@iamwhatever the push dismissed your approval again. Could you approve the workflow runs, re-approve once CI is green, and turn auto-merge back on? Thanks.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: onto main 699083f with no code change (the rebase applied cleanly).

The reds on 552db76 came from the base, not from this frontend-only diff:

  • Backend Tests (3.12, 4) and (Windows) (4) failed test_pid_reader_ratchet.py::test_pid_reader_sites_match_the_recorded_count_exactly: 160 sites against a recorded 158. Main fixed that count in fix(test): record the pid-reader count the tree actually holds #14778 (64b0bde) after this run's merge ref was built. Coverage Gate was red only because backend-test failed.
  • Cross-Platform Portability flagged added Python lines under src/kiro_crew/. This PR adds no Python, so those lines came from base drift in the merge ref.
  • Build Windows Installer (x64) hit the 120 s silent-install ceiling (test-windows-installer.ps1:221). That job passes on main's Build runs before and after this run.

tsc, eslint, the feature-map gate and 177 related frontend test files pass on the new head. @iamwhatever, could you please approve the fork workflow runs, re-approve once CI is green and re-enable auto-merge? Thanks.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Build Windows Installer (x64) failed on 1d7a58e (job) because the silent install went past the 120 s ceiling (test-windows-installer.ps1:221). The cause isn't this PR, which only changes frontend code. The wheel is the same size as passing PRs' (87.67 MB against 87.73 MB), the same job passed on other PRs at 66 to 86 s, and it failed the same way today on nine other PRs, including fix/instant-tip-touch, fix/composer-history-recall and feat/composer-list-indent.

I'm not pushing again. Could a maintainer rerun that job once the Opus review finishes? After CI is green, @iamwhatever, please re-approve and re-enable auto-merge. Thanks.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: main conflicted with this PR in website/src/pages/ChatSidebar.tsx. Main added the mc-touch-hit class to the mobile row's more-options button. This PR builds that button inside the memoized rowActions, so I added the same class there. Nothing else changed.

Before pushing, tsc, the feature-map gate and 178 related frontend test files (1,861 tests) passed, and so did eslint on this PR's files. The earlier Build Windows Installer failure was the repo-wide 120 s install timeout, and this push re-runs it.

@iamwhatever, could you approve the fork's workflow runs, re-approve once CI is green, and turn auto-merge back on? Thanks.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

CI on 1e70037: all 90 checks pass except the fork Opus 5 Review, which ended "review incomplete" after 33 minutes and posted no verdict (the Opus comment still shows the previous head's "no blocking findings"). PR Readiness is red only because of that lane. The code change since the last full Opus pass is the one-class mc-touch-hit carry-over from the rebase.

@iamwhatever, could you re-run the Fork Opus 5 Review for this head, then re-approve and turn auto-merge back on once readiness passes? I'm holding pushes so the approvals aren't dismissed again.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: main conflicted with this PR in docs/feature-map/README.md. I kept main's Sessions row and re-added this PR's pages/chat/useStableCallbackProps.ts entry; that entry is the only difference from main's text. No code change. The new head is 93516b3 on main 3a77091.

Before pushing, tsc, the feature-map gate and 178 related frontend test files passed, and so did eslint on this PR's files. This push also re-runs the Opus 5 review that ended "review incomplete" on 1e70037.

@iamwhatever, could you approve the fork's workflow runs, re-approve once CI is green, and turn auto-merge back on? Thanks.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

CI on 93516b3: 90 checks pass, including the Opus 5 review. The one failure is Backend Tests (Windows) (5), where a single test setup errored with OSError: [WinError 10055] (the runner ran out of socket buffer space) in test_member_prompt.py::TestMemberRulesRoutes::test_lone_surrogate_rules_are_400_not_500; 18,429 tests in that shard passed. This PR changes only frontend code and docs, so this is a runner flake. PR Readiness is red only because of that job.

@iamwhatever, could you re-run the failed job in run 36534138238, then re-approve and turn auto-merge back on once readiness passes? I'm holding pushes so the approval isn't dismissed again.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased: main conflicted with this PR in docs/feature-map/README.md again. I kept main's Sessions row and re-added this PR's pages/chat/useStableCallbackProps.ts entry, which is the only difference from main's text. No code change. The new head is e082638 on main 48b040c.

Before pushing, tsc, the feature-map gate and 178 related frontend test files passed, and so did eslint on this PR's files. This push also re-runs the Windows backend shard that hit the socket-buffer flake on 93516b3.

@iamwhatever, could you approve the fork's workflow runs, re-approve once CI is green, and turn auto-merge back on? Thanks.

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

CI on e082638: 91 checks pass. The one failure is Backend Tests (Windows) (1), with two tests in test/test_apps_registry_coverage.py:

  • TestSubdirGates::test_contained_join_degrades_to_none_on_a_symlink_loop
  • TestRunAppBuild::test_a_self_pointing_directory_link_under_a_deep_declared_path_is_judged_in_bounded_time[60]

This PR changes only frontend code and docs. The same two tests fail in the same shard on an unrelated PR (feat/composer-prompt-length-gate, run 36544307737), so the break looks like it came from main's recent apps changes (#13568 landed at 07:32Z) rather than from this diff. Main's push CI skips the Windows shards, which is why main itself shows green. PR Readiness is red only because of that shard.

I'm not pushing. Once main's Windows symlink tests are fixed, I'll rebase if needed; otherwise a rerun of that job after the fix is all this PR needs. @iamwhatever, when it's green, could you re-approve and turn auto-merge back on?

@Pearcekieser

Copy link
Copy Markdown
Contributor Author

Rebased onto current main to pick up #14986, which makes the two test_apps_registry_coverage.py symlink tests platform-aware. Those were the only failing tests on the previous head (Backend Tests (Windows) (1)), and they came from main, not this diff. The rebase applied cleanly with no code change; tsc, the feature-map gate and the targeted frontend tests pass locally.

@iamwhatever, could you approve the fork workflow runs, then re-approve and re-enable auto-merge once CI is green? Thanks.

@bolichen97 bolichen97 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Reviewed on behalf of the maintainers (sweep 2026-09-29): readiness passed, required checks green on d870e8f, no blocking review threads, security posture unchanged. Approving and merging.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

fork Pull request from a fork (external contributor)

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants