Repository navigation
plat 626
| Field | Value |
|---|---|
| State | open |
| Priority | P2 |
| Product | app |
| Area | tests |
| Summary | Current main has schedule/help/view/Vault test failures and undeclared authenticated-user prompt sections, separately from the fixed Relay landing assertion. |
Audit of dc4669ae0 from an isolated worktree, after fixing the Relay landing
assertion (PLAT-627):
-
SchedulesPage.test.tsx: four Back navigation tests expectsetShowWorkflowsOverview(false); the schedule component was unchanged by the Python Relay commit. -
WorkspacePanelGuideButton.test.tsx: two Skills help assertions expect the former answer count and “Skills for this project” text. This component was unchanged by the Relay commit. -
workspaceToolbarPlacement.test.ts: one assertion expects the old literalloadLegacyWorkspaceViewByPreset()[presetId] ?? 'report', changed in PLAT-551 (fea0bc976), before the Python Relay change. PLAT-577 did not touch it. -
GatewaySurface.test.tsx: fourteen failures readingstate.activeSessionsCache.findinChatTabPill. The shared tab status change is separate from the Relay patch; the Vault test store lacks that array. -
TestAgentWorksProductSurfaceE2EandTestCrewProductSurfaceE2Efail because the sharedauthenticated-usersection is absent from their declarations. That section was added by9b345dc75, not by PLAT-611.
Full frontend run: 4 failed / 504 passed / 1 skipped files; 21 failed / 2866 passed / 1 skipped tests. Both product declaration failures reproduced locally. Latest failed CI evidence: Frontend CI, Coding CLI P0. The frontend CI run precedes the new Vault tab-status failures and also includes the Relay assertion now fixed in PLAT-627.
P0 on the parent of PLAT-611 was already failing on the sandbox/secrets generated index; the PLAT-611 run had the same index failure. Current ticket checks pass. Parent P0.
None in this audit: the requested scope was failures caused by the Relay change. Existing frontend CI cleanup was previously recorded under PLAT-481; PLAT-571 also notes the stale workflow assertions, but neither is a current source of truth for this complete failure list.
Repair the affected fixtures/contracts after confirming their intended product behavior, run the named suites, then rerun frontend and P0 CI. Do not describe main as globally green until those checks pass.
Auto-synced from docs/ on main. Edit there, not here.