Repository navigation
test(app-sdk): press the queue control on busy slots in the optimistic-send test - #20212
Conversation
…c-send test
ChatEmbed mounts the shared ComposerSend atom. While a turn runs and the
editor has text, that control is named "Queue message", not "Send
message", so the three busy-slot cases in
ChatEmbed.optimisticSend.test.tsx found no button and the shard was red
on every frontend PR. The sendText helper takes { queue: true } and
clicks the queue control by role and name; the optimistic-row assertions
are unchanged.
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
UX Review (Opus 5.5) — ✅ PASSUX-level review of UX-Verdict: PASS Nothing to check. [UX-REVIEWED] 707462b Description read[DESCRIPTION-READ] 0083e4a4f8f307a267331c2d5e303f8da2a5c515da851a8d3fc96167bffb16b0 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 Nothing to check. [DESIGN-REVIEWED] 707462b Description read[DESCRIPTION-READ] 1855c8f45f7ab2a288c404374473c2bb2bca6d4e9a033a8d19de2897e33f5abd 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 Review detailsTest-only change, nothing to validate from the candidate list. No findings. [OPUS-REVIEWED] 707462b Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
KiroCrewCaptain
left a comment
There was a problem hiding this comment.
Tier 1 auto-approve: test (1 file). Criteria: no conflict, no requested changes, security path denylist clean, design-doc gate clean, security checklist all-NO, no reviewer-lane findings, readiness signal fresh for this head SHA, declared category matches the diff. Category: test-only change; the optimistic-send test now presses the queue control on a busy slot.
…rors
Bug A: stt_stream.py completes the WebSocket upgrade and sends
{type:error,code:stt_streaming_disabled} when streaming is off.
Client-side fix in useMeetingTranscription.ts:
- setActive(true) is deferred to the 'ready' frame handler so cleanup()
cannot flip active false on an error frame causing a reconnect loop.
- A connecting state tracks the period between worklet-ready and 'ready'.
The session effect calls stop() when !mayOpen && connecting, preventing
buffered PCM from being dispatched after a pause-window.
- stoppingRef.current = true in the error frame handler stops onclose
from emitting a second disconnected overwrite.
- Buffer flush after readyPromise is guarded by !stoppingRef.current.
- Guard before setConnecting(true): if wsRef.current !== ws, the socket
closed during addModule; returning prevents connecting stuck true.
- transcriptionErrorMessage clears when transcription reconnects and when
status leaves 'active'.
Bug B: meeting_lifecycle.py 409 body includes blocking_meeting_title and
blocking_meeting_status. api.ts extracts both into MeetingsApiError.
failureNotice picks between three i18n keys based on status.
MeetingsPage.tsx: no change (useNotifier reverted to original form).
ChatEmbed.optimisticSend.test.tsx: fix regression from PR #20212 which
changed getByLabelText('Send message') to getByRole for the non-queue
case. happy-dom's accessible name computation for aria-label-only buttons
differs from JSDOM; getByRole('button', {name:'Send message'}) fails in
that environment. Reverted to getByLabelText for the queue=false path.
Three new i18n keys in all 12 locale files. Source assertions updated.
Six screenshots committed.
Problem / Motivation
Goal: Fixes red main:
Frontend Tests (2)passes again onorigin/main, so frontend PRs stop going red on a test they did not touch.website/src/test/ChatEmbed.optimisticSend.test.tsxfails 5 of 12 cases on a cleanorigin/maincheckout (30bb0b3) withUnable to find a label with the text of: Send message. Failing job: https://github.com/kirodotdev/KiroCrew/actions/runs/38532361051/job/115749654088Why it matters
Every open frontend PR inherits this red shard, so none of them can reach
readiness: passeduntil main is green again.Not a goal
What changed (motivation → approach → change)
Two PRs crossed. #19881 added this test, which presses the composer's send control by the label
Send message. #20102 then replaced ChatEmbed's own Send button with the sharedComposerSendatom. While a turn runs and the editor has text, that atom's accessible name isQueue message(i18n keycomponents.chatInput.queue_message), so the three busy-slot cases found no button. Two idle-slot cases pass alone and failed only from state left by a case that threw mid-send.The fix is test-only. The
sendTexthelper takes{ queue: true }; the three busy-slot cases pass it, and the helper then presses the button namedQueue messageby role. This also checks that a busy slot offers the queue control rather than a plain send, which the old helper did not. Idle-slot cases still pressSend message.Backwards compatibility
Compatible: test-only change under
website/src/test/; no production code moves.Removes nothing: the diff adds an option to a test helper and changes three call sites.
Tests
website/src/test/ChatEmbed.optimisticSend.test.tsx: all 12 cases pass; the three busy-slot cases (on a busy slot shows the send as one queue card...,keeps both queue cards in place...,does not bring the card back...) now press the queue control and also lock in that a running slot names itQueue message.ChatEmbed*test files: 108 passed locally on this head.--sequence.shuffle; 12/12 each run.tsc -p tsconfig.app.json,eslinton the file,check_comment_history.py,check_brand_name.py,test_flake_pattern_ratchet.pyand the repo semgrep rules all pass.Manual verification
N/A — unit coverage sufficient: the change is to a test helper and the failing cases themselves are the proof.
Screenshots / video
Why no screenshot: test-only change; no rendered output moves.
Related Issues
Fixes #20206
Pattern harvest
Not generalizable: two PRs that touched the same control crossed in the merge window; the second changed an accessible name the first's test relied on.
Checklist
feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)