Skip to content

Draft: stack 22 later review PRs on the integration - #2667

Merged
r3dbars merged 48 commits into
mainfrom
claude/integration-wave2-20261112
Nov 13, 2026
Merged

r3dbars merged 48 commits into
mainfrom
claude/integration-wave2-20261112

Conversation

@r3dbars

@r3dbars r3dbars commented Nov 12, 2026

Copy link
Copy Markdown
Owner

What this is

The second batch of review PRs, stacked on the #2530 integration (e66e3e4). 106 PRs (#2558–#2663) were opened after #2530 was frozen. Each one was triaged against the integration head and then got an independent second read. 83 turned out to be already inside #2530 or to be notes that only record a review finding, so they aren't here. This branch carries the rest.

Included (22 PRs)

Merged with --no-ff, so each PR's exact head is an ancestor: #2558, #2561, #2568, #2570, #2571, #2574, #2605, #2606, #2609, #2610, #2613, #2614, #2618, #2619, #2626, #2634, #2643, #2645, #2647, #2659, #2662.

#2646 landed as its child commit only (632b19bf7, cherry-picked with -x): Stop and recovery read the engine, input node and pinned refs under the audio graph lock. Its parent #2639 is not on this branch.

What they do, roughly:

Conflict resolutions and follow-up commits

Not included

Checks (on this Mac, by the integrating agent)

  • bash check.sh quick: PASS (113 passed).
  • bash run-tests.sh: 54,340 of 54,341 passed before the pin fix; the one failure was the pin above. After the fix only the affected filter was rerun (85/85). The full fast suite was not rerun after that one-line test change.
  • bash build.sh --no-open: PASS.
  • TRANSCRIPTED_DISABLE_FILE_LOGGER=1 swift test: one failure in 714 Speaker tests (testInventoryCountsEachPersonsClipsMeetingsAndSpeech, fixture cleanup) while a build ran alongside. It passed 17/17 when rerun alone. Not quarantined and not fixed; none of these PRs touch it.
  • Source pins, module boundaries, file size, telemetry keys, banned calls: PASS.

Four area reviewers read the combined diff. Three found nothing; the fourth found the matrix rule above, now fixed.

Not verified: anything on real audio hardware, and CI on this exact head (dispatched separately).

🤖 Generated with Claude Code

…(UI-OVERLAY-A-8)

# Conflicts:
#	Sources/UI/Overlay/FloatingOverlayController.swift
… values (APP-DICTATION-7)

# Conflicts:
#	Sources/App/TranscriptedSupportActions.swift
… (UI-SETTINGS-A-4)

# Conflicts:
#	Sources/UI/Settings/AGENTS.md
#	Sources/UI/Settings/HomePresentation.swift
#	Tests/HomePresentationTests.swift
#	docs/repo-layout.md
# Conflicts:
#	Tests/TranscriptedWritingTests/Runtime/TildeSettingsTests.swift
PR #2613 deleted HomeFeedbackSheet, the only SettingsInlineActionButton call in the Home surface files, so the UI automation contract pin on it no longer holds.
The FileHandle rule now reads TranscriptFrontmatterScanner.swift, so the test-matrix trigger paths follow it. No rule reads the RetroactiveSpeakerUpdater files any more.
Base automatically changed from codex/integration-open-prs-20261110 to main November 12, 2026 22:13
The unannotated tuple from withAudioGraphLock compiled on Swift 6.4 but
not on the hosted CI toolchain, which inferred non-optional elements and
rejected the if-let bindings that follow (app-build, spm-tests and the
full bench all failed in Build dependencies). Use an explicitly typed
closure, as the Stop path in the same change already does.
@r3dbars
r3dbars merged commit 93b4e45 into main Nov 13, 2026
22 checks passed
@r3dbars
r3dbars deleted the claude/integration-wave2-20261112 branch November 13, 2026 10:21
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant