Skip to content

Commit 34aab3c

Browse files
committed
ci: run the tmpdir model tests off the Coverage lane
check:tmpdir-leaks:test proves the TMPDIR lifecycle by spawning a real nested `vitest run`. Sitting it in the Coverage job — seconds before the full instrumented suite — is a contention risk for no benefit, so it moves to the Layering Guard job. The Coverage lane already loses runs to `[vitest-pool]: Worker forks emitted error` when a fork is slow to terminate; main hit exactly that at 18291ba (scripts/fuzz/corpus-replay.test.ts, 708/709 files passed, zero failed tests) before this gate existed. That flake is not caused by this branch, but starting a second Vitest right before the big one can only make it likelier. The leak detector itself (`check:tmpdir-leaks`) stays in Coverage, where it has to run: it reads what the suite left behind, and it spawns nothing. Ownership is proven per unit of work, so the manifest is satisfied by any lane running these tests — moving the step keeps the gate wired. Co-Authored-By: Claude Opus 5 <[email protected]> Claude-Session: https://claude.ai/code/session_01SkS4S8XXrfkJ8TD1VBKkvJ
1 parent c6e42f4 commit 34aab3c

1 file changed

Lines changed: 14 additions & 7 deletions

File tree

‎.github/workflows/ci.yml‎

Lines changed: 14 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -116,6 +116,20 @@ jobs:
116116
# rather than in its own job so the two can never be green independently.
117117
run: pnpm depgraph:test
118118

119+
# The TMPDIR redirection both test lanes rely on (#1593/#1595). Its own tests ran
120+
# in no CI job at all until the #1429 gate manifest found them unowned.
121+
#
122+
# Deliberately NOT in the Coverage job next to `check:tmpdir-leaks`, where the
123+
# subject matter would put it: vitest-tmpdir-global-setup.test.ts proves the
124+
# lifecycle by spawning a real nested `vitest run`, and the Coverage lane already
125+
# loses runs to `[vitest-pool]: Worker forks emitted error` when a fork is slow to
126+
# terminate (observed on main at 18291ba8, unrelated to this gate). Starting a
127+
# nested Vitest seconds before the full instrumented suite is a contention risk
128+
# with nothing to gain — the manifest proves ownership per unit of work, so any
129+
# lane satisfies it.
130+
- name: Check the tmpdir redirection model
131+
run: pnpm check:tmpdir-leaks:test
132+
119133
affected-selector:
120134
name: Affected-check Selector
121135
runs-on: ubuntu-latest
@@ -266,13 +280,6 @@ jobs:
266280
- name: Check contention retry policy
267281
run: pnpm check:contention-retry
268282

269-
# The TMPDIR redirection both test lanes rely on (#1593/#1595). Its own
270-
# tests ran in no CI job until the #1429 gate manifest found them
271-
# unowned; they belong here, next to the other policy self-tests that
272-
# must pass before the suite they govern runs.
273-
- name: Check the tmpdir redirection model
274-
run: pnpm check:tmpdir-leaks:test
275-
276283
# Wrapped in the single-retry policy (#1419): a timeout-shaped failure in
277284
# an enumerated contention-flaky file reruns that file once and reports it
278285
# in the job summary. Assertion failures fail here on the first run.

0 commit comments

Comments
 (0)