Repository navigation
Merge upstream 62d98096ca0d into atli - #26
Conversation
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
… wrong folder (pingdotgg#18549) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
pingdotgg#20088) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…g#20980) Co-authored-by: Paul <paul@local>
pingdotgg#20069) Co-authored-by: Aaron Queen <bompus@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…pingdotgg#19937) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ingdotgg#19019) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…17922) Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
…ed to merge (pingdotgg#18232) Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…#18832) Co-authored-by: SAPHID <saphid@users.noreply.github.com> Co-authored-by: GPT-6.1 Sol <noreply@openai.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
… attachment (pingdotgg#21101) Co-authored-by: PR Batch Tester <agent@local.test> Co-authored-by: Tejas Rahane <17571323+trahane@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ingdotgg#14786) Co-authored-by: SAPHID <saphid@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ork advice (pingdotgg#17569) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…sh token (pingdotgg#21134) Co-authored-by: Aidid Haiqal <syafdo02@gmail.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…no origin (pingdotgg#20412) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…pingdotgg#20889) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ingdotgg#16026) Co-authored-by: Arthur Katcher <192321283+arthurkatcher@users.noreply.github.com>
…9525) Co-authored-by: Simone <185146821+Lucenx9@users.noreply.github.com>
…g#18647) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…report (pingdotgg#15738) Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
…ingdotgg#19366) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
…g#14411) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…ers (pingdotgg#13799) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…ingdotgg#13359) Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
…pingdotgg#21481) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
…gdotgg#13903) Co-authored-by: Julius Marminge <julius0216@outlook.com>
…ingdotgg#15044) Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
…14284) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…ngdotgg#18423) Co-authored-by: GPT-6.1 Sol <noreply@openai.com> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
…otgg#20349) Co-authored-by: Julius Marminge <julius0216@outlook.com>
…tgg#16040) Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Co-authored-by: SAPHID <saphid@users.noreply.github.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
… without permission (pingdotgg#20237) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…ingdotgg#19648) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
…ingdotgg#19968) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…nge (pingdotgg#11013) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
…ill (pingdotgg#20051) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…uanian, Basque or Azerbaijani locale (pingdotgg#19616) Co-authored-by: PR Batch Tester <agent@local.test> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Julius Marminge <julius0216@outlook.com>
… provider (pingdotgg#18467) Co-authored-by: Julius Marminge <julius0216@outlook.com> Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
…le can resume one (pingdotgg#21473) Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Upstream now releases a held queue when the user sends a message from a client (pingdotgg#21473). The provider-switch scenarios change the selection while the queue is held, so that message now comes over MCP, which keeps the hold. Delivered runs are found by id because they move after the message that ran ahead. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Upstream's web tests now run a headless Chromium project through Playwright. Fork CI's fork check and the full suite's web shard never installed that browser, so any web browser test failed to launch. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Skald review statusSkald: findings need a response (failure) at review state version 30, head Blocking
ClearedNothing is cleared at this epoch. Blocking: 0. Advisory: 3. Cleared: 0. Fenced: 0. Review rounds
|
There was a problem hiding this comment.
Skald review findings
Reviewed head 063ecb3326f8dba96a66ec4210cde4924241fd86 against base atli (4f69c29e4348cfeeedd2c5ae44af7a66743834d4).
New 1 · Carried 2 · Cleared 0 · Verified 0 since round 063ecb3
Not shown inline
These findings are listed here rather than on a line of this pull request: some have no line to attach to, and advisory ones are never threaded. Each row says which.
packages/client-runtime/src/pullRequestQuery.ts:70— Inherited Object.prototype members pass the undefined guard and are assigned as filter values even though they are not supported enum values. The query token is also removed from the remaining text. — no source location
Standing summary: open
Review details
Completed reviewers:
- architecture (hub-dispatch-v2:codex:gpt-6.1-sol:medium:9441865e1d0f674b)
- architecture (hub-dispatch-v2:codex:gpt-6.1-sol:medium:9441865e1d0f674b)
- architecture (hub-dispatch-v2:codex:gpt-6.1-sol:medium:9441865e1d0f674b)
- frontend (hub-dispatch-v2:codex:gpt-6.1-sol:medium:9441865e1d0f674b)
- security (hub-dispatch-v2:codex:gpt-6.1-sol:medium:9441865e1d0f674b)
- testing (hub-dispatch-v2:codex:gpt-6.1-sol:medium:9441865e1d0f674b)
- testing (hub-dispatch-v2:codex:gpt-6.1-sol:medium:9441865e1d0f674b)
- testing (hub-dispatch-v2:codex:gpt-6.1-sol:medium:9441865e1d0f674b)
| <NativeSegmentedControl | ||
| labels={props.options.map((option) => option.label)} | ||
| selectedIndex={props.options.findIndex((option) => option.value === props.selected)} |
There was a problem hiding this comment.
The new iOS implementation sends only option.label to UISegmentedControl and drops option.accessibilityLabel; the previous shared implementation applied that label to each accessible press target.
This finding is advisory tier: a code fix is optional and it does not block the merge on one, but an answer is required before merge — addressed, disputed, deferred, or duplicate.
Merges upstream commit 62d9809 (761 commits after #22). Merge with a merge commit.
This stops one commit short of upstream's tip,
1451e7cb23(in-app browser search engine). That commit passed CI, but it has collected 1,587 check suites while it is upstream's tip, and GitHub's commit check-runs API only returns the newest 1,000.scripts/fork-check.tstherefore can't see itsCheckand refuses it. The next update picks it up.Conflict resolutions:
pnpm-workspace.yaml/patches/expo-widgets@58.0.20.patch: upstream moved to expo-widgets 58.0.20, which already importsio.github.expo.peek, so the fork's fix(mobile): expo-widgets 58.0.18 compiles on Android #25 import rename is dropped and upstream's patch is used unchanged. The shiki-enginefbjnipatch stays (still React Native 0.88.0, shiki-engine 0.3.12). Upstream's@expo/ui58.0.20 and@expo/fingerprintentries are kept.NativeSwitch.android.tsx: upstream fixed switches inside pressable rows its own way (fix(mobile): prevent environment row taps from swallowing switches pingdotgg/t3code#17677: the Pressable owns the touch). The file now matches upstream, and the fork's responder hunk is dropped.ScheduledTaskService.ts: upstream fix(server): thread-bound scheduled tasks run on the thread's current provider pingdotgg/t3code#18467 makes the same fix as the fork (bound runs follow the thread's current provider), so this is upstream's version. The fork's "Uses this thread's current provider and model" settings row andfollowsThreadModelSelectionstay, in web and mobile.AdaptiveWorkspaceLayout.tsx,useThreadHeaderOptions.tsx,ThreadPullRequestsPanel.tsx,ScheduledTasksSettings.tsx,SettingsScheduledTasksRouteScreen.tsx,NewTaskDraftScreen.tsx: upstream moved these to the React Compiler (no manualuseMemo/useCallback). Each takes upstream's version with the fork's additions re-applied in that style. The search button upstream removed from the duo header is gone.ThreadHeader.tsx: the fork's Android thread-actions menu keeps the git and merge-back actions, and uses upstream's new merge glyph.schedule_taskdescription, ACP and OpenCode adapters: both sides kept (upstream's untrusted-webhook sentence and slash-command check, plus the fork's PR-watch flag).selfUpdate.test.ts,runtimeLayer.test.ts,orchestrationV2PendingBackgroundWork.test.ts: both sides' tests kept.pnpm-lock.yaml: upstream's, regenerated withvp i. It differs from upstream only by the fork's three patches (effect, react-native-image-viewing, react-native-shiki-engine).Fork code that needed changes for upstream (it merged cleanly but broke typecheck or tests):
ThreadRouteScreen: upstream dropped itsuseMemo/useCallbackimports, so the fork's thread-action handlers are plain functions now.runtimeLayer.test.tsprovideServerSettings, which the fork's watch setting requires.ProviderSwitch.integration.test.ts(separate commit): upstream fix(server,mobile): a message you send resumes a held queue, and mobile can resume one pingdotgg/t3code#21473 releases a held queue when the user sends a message from a client. The provider-switch scenarios change the selection while the queue is held, so that message now comes over MCP, which keeps the hold. Delivered runs are looked up by id because they move after the message that ran ahead.fork-ci.ymlandfork-full-suite.yml(separate commit): upstream's web tests now run a headless Chromium project through Playwright 1.60.0, which neither fork workflow installed. Fork CI's first run on this PR failed only because that browser was missing. Both now install it the way upstream's CI does (playwright install --only-shell chromium); the full suite needs it too, or the next release's web shard would fail.Verified locally: typecheck for mobile, web, server, desktop, client-runtime, contracts, shared, provider-acp and provider-opencode; lint and format on every hand-resolved file;
node scripts/fork-check.tspasses (every fork test file in mobile, server, web, client-runtime, contracts, provider-core, shared and scripts). No upstream workflow was added, so none needs disabling.Done by Claude Opus 5.5 in Claude Code (T3 Code).
🤖 Generated with Claude Code