Carlos Florêncio
2.5 KiB
2.5 KiB
id, title, status, wave, depends_on, plan, spec
| id | title | status | wave | depends_on | plan | spec | |
|---|---|---|---|---|---|---|---|
| 11-e2e-temp-cleanup | Contain E2E temporary roots | complete | 7 |
|
plan.md | ../../specs/system-page/requirements/storage-maintenance.md |
Task 11: Contain E2E temporary roots
Prevent normal managed E2E runs from leaving large numbers of kandev-e2e-* backend roots in the
service's inherited temporary directory.
Acceptance
- Backend fixture directories created by one managed run have one explicit run owner and are removed on normal success and failure without deleting another concurrent run's files.
- The cleanup guard begins immediately after
mkdtempSync, covers setup/spawn/initial health failures, passes the spawned process to the initial health wait for fast early-exit detection, and reports cleanup failures instead of silently discarding them. - Cleanup targets only the exact run-owned root; interrupted-process residue remains subject to host temp policy unless a safe existing runner lifecycle can prove ownership.
- Tests or a deterministic harness regression prove success/failure cleanup and concurrent-run isolation. No test uses the host's real Kandev data directory.
Verification
cd apps/web && pnpm e2e:run tests/system/storage-maintenance.spec.ts
cd apps/web && pnpm e2e:run tests/system/mobile-storage-maintenance.spec.ts -- --project=mobile-chrome
Files likely touched
apps/web/e2e/fixtures/backend.ts- a focused fixture-lifecycle test beside the existing E2E harness tests
- focused fixture tests
Dependencies
Task 10.
Inputs
- Read-only audit of 1,099 abandoned
kandev-e2e-*roots - Existing managed runner cleanup and concurrency conventions
Output contract
Report root ownership, success/failure/interruption semantics, concurrency safety, tests/commands, remaining host-policy boundary, and update this task plus the serialized plan checkbox when done.
Completion
- The atomically-created worker root is now guarded immediately after creation. Setup, spawn, initial-health, test-use, and restart failures all converge on the same exact-root cleanup path.
- The latest registered backend process group is stopped before bounded directory removal. Cleanup failures are surfaced, including alongside the original fixture failure.
- Focused lifecycle coverage proves early child-exit detection, success/failure cleanup, process-stop ordering, peer-root isolation, and cleanup-error reporting.
- Both required managed Storage E2E commands passed, and neither left a new
kandev-e2e-*root.