timothyandClaude Opus 5
66c8500e94
PR Gates / CI image pin matches docker/ci (pull_request) Successful in 17s
PR Gates / Docs update reminder (pull_request) Successful in 23s
PR Gates / decisions lifecycle (pull_request) Successful in 25s
Review verdict / Set review-verdict status (pull_request) Successful in 7s
PR Gates / Script tests (pytest) (pull_request) Successful in 46s
Build ErsatzTV Image / API docs in sync (OpenAPI + endpoint index) (pull_request) Successful in 18s
Build ErsatzTV Image / Formatting (changed .cs conform to .editorconfig) (pull_request) Successful in 15s
Build ErsatzTV Image / Functional E2E (curl + UI contracts) (pull_request) Successful in 5m51s
Build ErsatzTV Image / EF migration integrity (SQLite + MySql) (pull_request) Successful in 19m40s
Build ErsatzTV Image / Build & test (.NET) (pull_request) Successful in 21m57s
Build ErsatzTV Image / Build & push image (amd64) (pull_request) Has been skipped
review-verdict/h10 Review-verdict: MERGEABLE @ 66c8500 (base: main)
FIX 1: `PlaylistsScreen.test.tsx`'s group-id test filtered POSTs and asserted inside a `for` loop over the result. On the fixed build that array is empty — the unmatched value leaves the select at '', so Create is disabled and jsdom won't dispatch its onClick — so ZERO assertions ran and the title claimed more than the body proved. A future change that re-enabled Create and POSTed `playlistGroupId: null` would still have passed. Added the unconditional `expect(posts).toHaveLength(0)` before the loop. Worth noting this is the ninth instance of the shape on this branch, and the sibling strengthening in the *same commit* got it right (`RerunCollectionsScreen.test.tsx` uses an unconditional `toHaveLength(0)`) — so the lesson didn't generalize even one file over. The rule is: an assertion inside a loop over a filtered collection proves nothing until the collection's length is asserted. FIX 2: "1 item need a selection" — the noun was pluralized, the verb wasn't, and singular is the common case. My test used `/need a selection/i`, which matches both the right and wrong grammar, so nothing could catch it; it now asserts the exact string '1 item needs a selection'. FIX 3: two comments about clicking an already-disabled button read as contradictory policy. They're not — on the parent the Playlists button was ENABLED, so there the click genuinely discriminates, while the rerun button was disabled on both sides, making it a restatement of `toBeDisabled()`. Both comments now say which case they are and why. Added to #677: the row label falls back to "(no X selected)" on empty `selectedName` regardless of `selectedId` (the mirror image of the bug fixed here), and an all-unbindable group list disables Create with no reason shown. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>