test: SettingsView.leaveGuard-13163 "after a successful save the page is clean again" failed once on main's scheduled coverage shard 5/8 (run 39703234572), green on the same commit's push run #13317

Activity

1. objectstack-fleet

   added

   area:devpathThe road — create, dev, verify, publish/install, connect an agent, iterateThe road — create, dev, verify, publish/install, connect an agent, iterate

   bugSomething isn't workingSomething isn't working

   domain:uiobjectui ui stream: fix lands on the published library or apps — objectui execution seatobjectui ui stream: fix lands on the published library or apps — objectui execution seat

   pm:queue

   priority:p3

   pm:dispatched

   and removed

   pm:queue

   on Oct 22, 2026
2. objectstack-fleet

   assigned

   os-sales

   on Oct 22, 2026
3. objectstack-fleet commented on Oct 22, 2026

   

   objectstack-fleetBot

   on Oct 22, 2026

   ContributorAuthor

   More actions

   Claim: PM loop round 1
   Session: session_019Bd3muSBdHUFdJiKqnsM2W
   Account: os-sales (the seat's linked user as GET /user answers it; always the card's assignee)
   Branch: claude/issue-13317-settings-leaveguard-race
   Worktree: objectui-issue-13317
   Domain: domain:ui
   Seat: domain:ui#3
   File surface: apps/console/src/pages/settings/__tests__/SettingsView.leaveGuard-13163.test.tsx; apps/console/src/pages/settings/SettingsView.tsx only if the race is in the product code; .changeset/13317-*.md if product code moves (stop on breach; explain in the report)
   Container & model: M, mode:subagent, model: opus (default tier; dispatch-gates --tier --repo objectstack-ai/objectui: no path-derived mandate)
   Clause-②: no
   Responsibility: the test (or SettingsView's post-save dirty reset) in objectui apps/console | none | the scheduled coverage shard 5/8 on objectui main, once in 12 runs read; no user path is shown yet
   Thread-read: none
   Serial constraints cleared: none — no open PR or in-flight claim touches apps/console/src/pages/settings/** (last changes on objectui origin/main cd14d8e: 0093b2c objectui#13298 and 1ca759a objectui#13262, both merged)
4. objectstack-fleet commented on Oct 22, 2026

   

   objectstack-fleetBot

   on Oct 22, 2026

   ContributorAuthor

   More actions

   os-dev-report
   {
   "issue": 13317,
   "status": "done",
   "branch": "claude/issue-13317-settings-leaveguard-race",
   "pr": "#13322",
   "session": "session_019Bd3muSBdHUFdJiKqnsM2W",
   "premise_still_valid": true,
   "summary": "Race found and reproduced. The case waited for the unsaved bar to leave the DOM, which happens in the commit that renders setDraft({}) after the save, and then synchronously asserted that beforeunload is not held. The guard is released one step later, in the passive-effect cleanup of useRegisterUnsavedEdits. The save settles outside act(), so the render is on the default lane and react-dom 19.2.8 schedules the passive flush as a separate scheduler task (only sync-lane commits flush inline). When the commit task overruns the scheduler's 5ms frame the scheduler yields, and waitFor (whose wrapper ends on setTimeout(0)) can return before that task runs. Fix is test-only, at the expected landing site: wait on the guard itself (waitFor around the beforeunload check), the shape plugin-form's discardGuard.test.tsx already uses. Product code is unchanged because the release lands one task after the commit, before any paint a user could react to. PM assumption 1 holds with a correction: the draft reset lands on time and the guard release is the late step. Assumption 2 holds.",
   "tests": "Baseline at base cd14d8e: pnpm exec vitest run on the file, 12/12 passed (lock VERDICT command-exit 0). Uncommitted probe, deleted after use: when the bar leaves the DOM (MutationObserver) the guard was still held 20/20 on an unloaded run; right after waitFor it was released 20/20. With a slow machine modelled (performance.now skewed +10ms per read so React's scheduler yields after every task), the guard was still held right after waitFor 13/20. The real file under that same load, with the case repeated 20x (uncommitted copies, deleted): before = 'Tests 10 failed | 21 passed (31)', every failure 'AssertionError: expected true to be false' at expect(unload.defaultPrevented).toBe(false), the other 11 cases green; after = 'Tests 31 passed (31)'. Ablation, committed fix first: node objectstack/scripts/ablation-replace.mjs mutated useUnsavedLeaveGuard.ts so that useRegisterUnsavedEdits(dirty, message) became useRegisterUnsavedEdits(true, message) (anchor 1 to 0, blob 42a45f17ddd5 to 97ee57239286). The fixed case went red inside the new waitFor ('Tests 1 failed | 11 skipped', expected true to be false). Restored: blob == HEAD 42a45f17ddd5, git diff HEAD empty. No dist is involved (the console app resolves from source), so no rebuild was needed. At head 42cc11c: pnpm exec vitest run apps/console/ gave 'Test Files 236 passed (236)', 'Tests 2408 passed (2408)' (lock VERDICT command-exit 0).",
   "gates": "objectui families were derived by hand from objectui package.json and workflows (objectstack dispatch-gates refuses objectui). Local, head 42cc11c, each exit code captured before any pipe and written to a file: node scripts/check-changeset-presence.mjs 0 (1 empty-frontmatter changeset declared); node scripts/check-changeset-no-major.mjs 0; eslint on the changed file with the console config 0; pnpm check:new-line-citations 0; pnpm check:control-bytes 0; pnpm check:test-path-roots 0; pnpm check:changeset-claims 0; pnpm check:pending-changeset-literals 0; vitest apps/console (repo-root form) 0. NOT MEASURED: pnpm --filter @object-ui/console type-check exit 2, because the worktree has no built workspace dist (821 errors, all TS2307/TS2882 module resolution or cascades from it; the changed file's only error is its existing @object-ui/i18n import). Repository-wide pnpm lint was not run locally (CI owns it). PR check-runs at report time, 35 total: Changeset Fixed Group Check success, label success; Test (coverage) and the coverage shard matrix skipped (they run on schedule only, so this PR never exercises the condition that flaked); Type Check, Lint, Test shards 1-16/16, Test (dist pins), Repo Gates, Line Citation Gate, Changeset Declaration, Changeset Claim Re-read, Changeset Bump Policy, Governed Surface Queue Guard, Build and E2E, Build Docs, Bundle Analysis, Changeset Overwrite Report, Live E2E (informational) in_progress.",
   "dogfood_battery": "n/a: the PR title has no '!' and the claim's Clause-② is no",
   "line_budget": "n/a: no skills/** or ratcheted ledger touched",
   "files_changed": [
   "apps/console/src/pages/settings/tests/SettingsView.leaveGuard-13163.test.tsx",
   ".changeset/13317-settings-leaveguard-race.md"
   ],
   "deviations": [
   "The dispatch asks for pnpm lint and pnpm type-check locally, but os-dev.md assigns repository-wide lint to CI. I ran eslint on the changed file and left pnpm lint to CI. The local console type-check is NOT MEASURED because the workspace dist is not built, and building the console's dependency closure was judged out of proportion for a test-only diff, so CI's Type Check decides it.",
   "The claim's file surface lists .changeset/13317-*.md only 'if product code moves'. I added an empty-frontmatter changeset anyway: objectui's changeset-presence gate counts any file under apps/console/src, tests included, and the dispatch says a test-only fix follows AGENTS.md section 9, which asks for exactly this empty declaration.",
   "Commit trailers use objectui AGENTS.md's model-free pair (Claude-Session plus Co-authored-by: Claude), not the harness's model-named trailer, and the PR body ends with the os-dev session-URL footer instead of the harness's footer. Both follow the precedence rule."
   ],
   "mcp_calls": "0",
   "api_writes": "2 relay dispatches: (1) pr_create = POST /repos/objectstack-ai/objectui/pulls (draft) + POST /repos//issues/13322/assignees [os-sales]; (2) os-dev-report comment = POST /repos//issues/13317/comments. git push of the branch (2 pushes: empty probe, then commits). No label writes: none named by the dispatch, and objectui's skip-changeset has no readers.",
   "open_questions": [],
   "out_of_scope_findings": [
   "carrier: none. Noted, not filed. Survey of the other suites that assert beforeunload (git grep -ln beforeunload over tests): none waits on a DOM stand-in for a guard that is released after an awaited save, so no sibling of this shape was found. Dedupe words: leave guard passive effect waitFor beforeunload flaky"
   ]
   }

   ---

   Generated by Claude Code
5. objectstack-fleet commented on Oct 22, 2026

   

   objectstack-fleetBot

   on Oct 22, 2026

   ContributorAuthor

   More actions

   ACCEPT → PR #13322 comment 6338011186 · domain:ui seat 3 · session_019Bd3muSBdHUFdJiKqnsM2W · head 42cc11c · 2026-10-22T09:59Z. Lands once every check on that head is green.
6. objectstack-fleet

   closed this as completedin #13322on Oct 22, 2026
7. objectstack-fleet

   removed

   pm:dispatched

   on Oct 22, 2026
8. objectstack-fleet commented on Oct 22, 2026

   

   objectstack-fleetBot

   on Oct 22, 2026

   ContributorAuthor

   More actions

   Landed · domain:ui seat 3 · session_019Bd3muSBdHUFdJiKqnsM2W (os-sales) · 2026-10-22T10:24Z

   PR #13322 merged through the queue as 4c212b9e0 (2026-10-22T10:22Z). This card closed completed through its Fixes.

   reading (SettingsView.leaveGuard-13163.test.tsx)	base cd14d8e	origin/main 4c212b9
   the fix's objectui#13317 comment line	0	1
   control: beforeunload lines, same file	5	5
   .changeset/13317-settings-leaveguard-race.md	absent	present

   The after-reading for the flake itself is the scheduled coverage runs on main, which this PR does not exercise. A recurrence of this case after 4c212b9 is a new signature, not this card.

   State change in the same stroke: pm:dispatched removed.