Repository navigation
fix: backport 0.12.1 batch 63 (6 docs/tests-conflict fixes) - #32927
Conversation
Escape typed into an editable field of a side-panel file tab (the code editor, or the Files rail's filter) no longer closes the file. The tab listened for Escape on the whole page and the editor does not claim the key, so a press meant for the editor closed the file mid-edit. Escape from anywhere else, the chat composer included, still closes it. Closes #23975 (cherry picked from commit 27fce31) Original author: Bolin Chen (#31653). Hand-resolved, docs/tests only: docs/system-specs/modules/learn-cron-dashboard.md keep release's structure plus this commit's added and removed lines.
…31893) A restarted gateway accepts WebSocket connections before its boot restore has put the chat slots back. A dashboard tab that reconnects in that window sends subscribe_subagents for slots that are not open yet, so the replay queue refused every one of them and the Subagents panel stayed empty until the user focused another slot. An owner socket's replay queue now waits for the existing startup_tabs_rebuilt event (bounded at 300s) before its first batch, and admits the slots asked for meanwhile only after it. An app socket does not wait: its visible slots are read at subscribe. (cherry picked from commit a17bda3)
…#31442) The cron help, tool schema and docs spell a script as ~/.kiro/crew/crons/<file>.py, but the resolver expanded ~ against the real home, so under a KIROCREW_HOME override the documented spelling never resolved. Rebase that prefix onto config_dir() before resolution. A spec already inside the data home (an override nested under ~/.kiro/crew, compared on canonical paths) is kept, and the resolve, sensitive-path and containment checks still judge the result. Closes #29828 (cherry picked from commit 1694a06) Original author: Bolin Chen (#31442). Hand-resolved, docs/tests only: docs/system-specs/modules/learn-cron-dashboard.md keep release's structure plus this commit's added and removed lines.
…on (#31106) * fix(chat): bound a background slot's live-appended transcript retention * fix(chat): hydrate a split pane again once its slot stops being active A split pane that shows the active slot has its hydrate refused, because the active mirror is live. When the reader switches away while that view is empty or still loading, the park writes no page marker, so the pane's hold cannot flag the slot either. Nothing flagged it afterwards, and a finished background turn then trimmed rows the pane still showed. The pane's hydrate effect now also runs when the slot stops being active, so the page refused earlier is placed and the slot is flagged. --------- Co-authored-by: Gary Jennings <garysjennings@gmail.com> (cherry picked from commit db008d5) Original author: Pearcekieser (#31106). Hand-resolved, docs/tests only: website/src/store/chatSlice.publicSurface.test.ts keep release's structure plus this commit's added and removed lines.
On the Claude Code harness a message sent while a tool ran was refused and queued for the next turn, because a _session/steering request during a running tool makes claude-agent-acp background the tool and end the turn early. Nothing retried it when the tool ended, so a mid-turn message waited for the whole turn, and a wait sleep slept through it. Hold such a steer, stamp it so a wait sleep ends, and send it once the turn's last open tool call ends, oldest first. A held steer the turn never takes is queued by the existing teardown, marked possibly delivered when its write may have reached the adapter. Builds on #25000. Co-authored-by: Aaron Longwell <a@adl.io> (cherry picked from commit 13e2433) Original author: Bolin Chen (#32372). Hand-resolved, docs/tests only: docs/system-specs/modules/harness-parity.md keep release's structure plus this commit's added and removed lines.
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS Nothing blocks: each item is a fix that has already merged on What this change shipsInventory (8 items) — 8 justifiedIntent: bring six user-visible fixes that already merged on
[FIRST-PRINCIPLES-REVIEWED] 4767fe9 Description read[DESCRIPTION-READ] 19d5ebe7b44574aa3851bed580752b98d812e2193a3e40d439ed5dce16a2e59f That sha256 names the PR description this verdict read. Recompute it with |
UX Review (Opus 5.5) — ✅ PASSUX-level review of UX-Verdict: PASS Nothing blocks: the diff adds or changes no visible control, label or string, so nothing needed a screenshot. The one interaction change is Escape in the side-panel editor, which now leaves the open file open. Notes
[UX-REVIEWED] 4767fe9 Description read[DESCRIPTION-READ] ff7b9d23ac9c24ef3932d3b30c9c5bda2b1ffd59b61be72527efca25303e8d8d 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 |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS This is a straight cherry-pick of six fixes already merged on Notes
[DESIGN-REVIEWED] 4767fe9 Description read[DESCRIPTION-READ] 279dd86808ecac356619b79da2118e5e058605657bc8e6ebcedbefaaee6e4c63 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 |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
4629996 to
4767fe9
Compare
|
Goal edited at the maintainer's request: scope split between builders. This PR now carries only the value-2 docs/tests-conflict rows (#31653, #31893, #31442, #31655, #31106, #32372); the value-1 rows and #30084 move to another builder's batch, and #31400 is left out because an open release PR records that it needs a prerequisite. Head 4629996 -> 4767fe9. |
Problem / Motivation
Goal: Bring 6 small fixes that merged on
mainafterv0.12.0ontorelease/0.12.0, one commit each, in a PR of their own.Why it matters
Each row is a user-visible value-2 fix from the 0.12.1 wide-bar list (Ray, 11-07) that an earlier builder skipped because its pick conflicted in docs or tests only.
Not a goal
main; nothing is changed only here.CHANGELOG.mdchange.What changed (motivation → approach → change)
Each main squash commit was cherry-picked with
-xontorelease/0.12.0@3c6a160c8a, in merge order onmain, one commit each. #31893 applied cleanly. The other five conflicted in docs or test files only and are hand-resolved: keep release's text and structure, then apply the fix's own added and removed lines. No production code is hand-resolved, and no prerequisite is picked.docs/system-specs/modules/learn-cron-dashboard.mdwebsite/src/store/chatSlice.publicSurface.test.tsdocs/system-specs/modules/dashboard-token-auth.mddocs/system-specs/modules/harness-parity.md#31653, #31442 and #31655 have added and removed lines identical to the main commit. For #31106 and #32372, release words the line the fix rewrites differently, so the fix's word-level edit is applied to release's line; each differs from the main commit by exactly that one line.
27fce313cb5b9ff226faa17bda3d0889e895f5911694a06f92897a7ca997db008d59d43336dc97cac1924789e0db1ebdd51d13e24338b64767fe9d0e5 commits carried a non-public author email or an internal co-author trailer, which the push guard refuses, so they are authored with the public identity and name their original author in the message; internal trailer lines are removed. The other 1 keep their original author.
builder-security checked: each change is the one already merged on
main; no bind, listener, new route, auth mechanism or storage grant is added.Backwards compatibility
Compatible: each change is the one already merged on
main, and its compatibility note applies unchanged.Removes nothing: no setting, key or caller is removed.
Tests
Proof: not provable -- other: a backport of main commits whose proofs ran on their own PRs
release/*, so the direct test modules were run here: every touched test file plus every test file that imports a touched module, 166 files,-n 4 --dist loadgroup. 6367 passed, 4 skipped, 2 failed; both fail the same way on a cleanrelease/0.12.0@3c6a160c8aon this host.tsc -p tsconfig.app.json --noEmitpasses.vitest runon every touched test file and component name: 187 files, 3160 passed, 1 failed (MarkdownPanel.editFocus.test.tsx, which fails the same way on clean release).Manual verification
N/A - each fix was verified on its own main PR. The 0.12.1 RC soak is the end-to-end check.
Screenshots / video
Why no screenshot: each frontend pick carries the same rendered change as its main PR, whose screenshots show it.
Related Issues
Backport to the 0.12 line for 0.12.1 of #31653, #31893, #31442, #31106, #31655, #32372.
Sibling sweep
Siblings: none -- each source PR carried its own sibling sweep; this PR only carries those fixes.
Pattern harvest
Not generalizable: this PR only carries fixes already on
main. The source PRs have their own harvest.Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)Contribution License Agreement
Contributions are accepted under the project's Apache-2.0 license; see the Licensing section of CONTRIBUTING.md.