Repository navigation
fix: backport #32387, #32434 and #32833 to 0.12.1 - #32823
Conversation
…32387) A dashboard tab whose turn called spawn_sub_agents acknowledged each returned child on the mark, so a turn the user stopped before it landed lost the result. Every parent now holds the mark pending under the spawning attempt and settles it at that attempt's release, as cron and channel parents already do (#27508): a landed turn confirms it, any other end delivers it. The specs and the brokered-delivery RFC no longer describe the removed dashboard-born exception. Builds on #28576 by @cruisercohen. Closes #23495 (cherry picked from commit 2a4fb38) Original author: Bolin Chen (#32387). docs/system-specs/modules/subagent.md hand-resolved: removed the dashboard-born-tab sentence this PR makes untrue and kept the rest of release's paragraph; main's #25261 sentence belongs to a main change not on release.
UX Review (Opus 5.5) — ✅ PASS (a BLOCK with no Trigger and no Outcome does not block)UX-level review of UX-Verdict: PASS cannot evaluate: missing screenshots of the embedded chat panel's new "session gone" notice and its "Try again" failure state. The PR attaches no image, so the blind read was not performed. Evidence gaps
Notes
[UX-REVIEWED] 57288b4 Description read[DESCRIPTION-READ] fad95cf01463731f605d815dc32c4254b090948dc357e24144ca4533c7985b31 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) — ✅ PASS (a BLOCK with no Trigger and no Outcome does not block)Design-level review of Design-Verdict: PASS cannot evaluate: missing screenshot or recording of the embedded ChatPanel's "session gone" notice and the new failure/Try-again state in ChatPanel and ArtifactChatPanel, captured on this revision. BlockersThe reshaped embedded-chat surfaces have no rendered evidence from this PR Notes
[DESIGN-REVIEWED] 57288b4 Description read[DESCRIPTION-READ] 721b8dc8510777fc0ed3bbdb2d05b3d1060fa25455cdea2df085f39725b770a0 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 |
First Principles Review (Opus 5.5) — ✅ PASSPremise-level review of First-Principles-Verdict: PASS This backports three fixes already merged on What this change shipsInventory (8 items) — 8 justifiedIntent: let 0.12.x users get the fixes for lost subagent results and for embedded chat panels that take over or stick on the main chat's session. This is a FIX (a backport).
[FIRST-PRINCIPLES-REVIEWED] 57288b4 Description read[DESCRIPTION-READ] dfeb26738c4175041c717c4eb40bd9286fee52645cd4e27395176aba9c5c172b That sha256 names the PR description this verdict read. Recompute it with |
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: |
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: |
84e8c14 to
b4e7962
Compare
|
Goal edited at the maintainer's request: two fixes -> #32387 alone. #32434 is removed from this PR (head 84e8c14 -> b4e7962) because GPT 6.1's first-run finding on |
The artifact companion and the app SDK ChatPanel embed ChatPage, which renders the dashboard's active slot, and they switched to their session without waiting for the answer. A deleted session or a failed open left the previously open chat on screen, and messages typed in the panel went to that other session. Both now use useOwnSlotChat, the guard the Papyrus co-author panel has had since #22460: the chat shows only while the panel's own session is active, a 404 is reported to the host as a gone session, and any other failure shows an ErrorNotice with Try again. Switching back to the panel's session waits for another session's switch to finish and runs as a transition, so a control that opens another session and navigates away keeps the session it opened. Closes #25341 (cherry picked from commit 3fc6ab4) Original author: Bolin Chen (#32434). website/src/i18n/en.context.json hand-resolved: appended only this PR's appSdk.chatPanel.session_gone translator note after release's last entry; the notes between belong to main changes not on release.
An embedded chat panel opens its session with switchSlot and, on a 404, has the host forget that session. The 404 was acted on even when it arrived after the panel had closed, so a session the user reopened from History in the meantime was removed and switched away from. A late answer to a closed panel's open is now ignored, as a late success and a late failure already were. Refs #32434 (cherry picked from commit cfca397)
Problem / Motivation
Goal: Bring two user-reported fixes, #32387 and #32434 with #32434's follow-up #32833, that merged on
mainafterv0.12.0ontorelease/0.12.0, one commit each, in a PR of their own.Why it matters
Each fix closes a problem a user reported. Each faulty line was checked on
release/0.12.0@052ce87b11before the pick.dashboard/messaging_api/run_control.py:465website/src/components/ArtifactChatPanel.tsx:63(dispatch(switchSlot(slotKey)))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@052ce87b11, in merge order onmain.2a4fb38447b4e796281a3fc6ab4f91218c3908e0cfca397e0f57288b4191Hand-resolved hunks, docs only:
docs/system-specs/modules/subagent.md(one paragraph)website/src/i18n/en.context.jsonappSdk.chatPanel.session_gonetranslator note after release's last entry; the notes between belong to main changes not on release. The file parses as JSON with no duplicate key.Every other file has added and removed lines identical to its main commit.
#32833 is the
mainfix for the one finding GPT 6.1 raised on #32434 (website/src/hooks/useOwnSlotChat.ts:115): a "session not found" answer that lands after the panel unmounted no longer callsonSessionGone. It picked cleanly.All three commits are authored with the public identity (#32387 and #32434 are hand-resolved and #32387 carried a co-author trailer with an internal address; #32833's author email is not a public one), so they are authored with the public identity and name their original author (Bolin Chen) in the message; the trailer line is removed.
builder-security checked: #32387 changes when the gateway settles a returned result inside existing routes; no bind, listener, new route, auth mechanism or storage grant changes.
Backwards compatibility
Compatible: each change is the one already merged on
main, and its compatibility note applies unchanged. On a dashboard tab a returned result whose turn did not end onend_turnnow gets its completion turn (one duplicate at worst, never a loss), where before it was dropped.Reader: src/kiro_crew/parent_turns.py:settles_at_release -- chat -- removed by #32387; every caller (run_control, subagent_injection, and the gateway and messaging re-exports) is changed here, and git grep finds no other reader in src/ or test/ on this branch
Reader: src/kiro_crew/dashboard/messaging_api/run_control.py:api_spawn_mark_collected -- chat -- a dashboard tab's queued completion is no longer settled at mark time; the landed release settles it (test_collecting_a_member_whose_announce_already_queued_settles_it)
Reader: src/kiro_crew/slack/gateway_runtime/subagent_injection.py:_route_subagent_to_tab -- chat -- skips only a confirmed mark and parks a pending one (test_a_tab_parents_landed_turn_plays_no_duplicate_and_settles_at_its_release[dashboard-tab])
Tests
Proof: not provable -- other: a backport of main commits whose proofs ran on their own PRs
release/*, so the related set was run here: every touched test file, every test file that imports a touched module, and the tests that read the touched docs, 154 files,-n 4 --dist loadgroup. 5158 passed, 2 skipped, 1 xfailed, 0 failed.tsc -p tsconfig.app.json --noEmitpasses.vitest run ArtifactChatPanel AppSdk appSdk ChatPanel useOwnSlotChat switchSlotCallsiteClassification ArtifactDetailPage.companionChat i18n locale context: 151 files, 2635 passed, 0 failed.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: the embedded chat panel's "session gone" notice is the same rendered change as on the main PR, whose screenshots show it.
Related Issues
Backport to the 0.12 line for 0.12.1 of #32387, #32434 and #32833. Source issues: #23495, #25341.
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.