Skip to content

Revert "fix(test-runner): disallow skips in after hooks" - #43887

Merged
Pavel Feldman (pavelfeldman) merged 1 commit into
mainfrom
revert-43713-fix-43676
Nov 3, 2026
Merged

Pavel Feldman (pavelfeldman) merged 1 commit into
mainfrom
revert-43713-fix-43676

Conversation

@dgozman

Copy link
Copy Markdown
Collaborator

Reverts #43713
Considered a breaking change

@pavelfeldman
Pavel Feldman (pavelfeldman) merged commit 7e39e17 into main Nov 3, 2026
29 of 31 checks passed
@pavelfeldman
Pavel Feldman (pavelfeldman) deleted the revert-43713-fix-43676 branch November 3, 2026 18:57
@github-actions

github-actions Bot commented Nov 3, 2026

Copy link
Copy Markdown
Contributor

Test results for "MCP"

1 failed
❌ [firefox] › mcp/cli-session.spec.ts:148 › session stops when socket directory in TMPDIR disappears @mcp-ubuntu-latest-firefox

5193 passed, 616 skipped


Merge workflow run.

@github-actions

github-actions Bot commented Nov 3, 2026

Copy link
Copy Markdown
Contributor

Test results for "tests 1"

1 failed
❌ [default-reuse] › run-tests.spec.ts:1041 › should produce output twice @vscode-extension

23582 passed, 704 skipped


Merge workflow run.

@github-actions

github-actions Bot commented Nov 3, 2026

Copy link
Copy Markdown
Contributor

🟡 Failures look unrelated, but the VS Code failure remains uncertain

Hi, I'm the Playwright bot. I checked both CI failures for a355853 against the diff and independent runs.

Details

The PR only restores support for skips in after hooks and removes two tests for the reverted restriction. Neither failure points to that change.

Pre-existing flake

  • [firefox] › mcp/cli-session.spec.ts:148 › session stops when socket directory in TMPDIR disappears — the daemon is still alive after the five-second wait. The refreshed test-results database shows the exact same test and Ubuntu Firefox configuration failing with the same assertion on main at 47ba108, before this PR. Excluding this PR, it failed in 5 of 193 runs and passed in 188. This is an existing shutdown-timing flake, not evidence against the after-hook change.

Uncertain — likely unrelated infrastructure

  • [default-reuse] › run-tests.spec.ts:1041 › should produce output twice — the VS Code job fails because its expected output gains a GitCommitInfo: timeout ... git fetch origin db6633be... line. The inner test passes; the output comparison fails. Git metadata collection runs during setup, not in the changed after-hook code.
  • The same output mismatch appears on independent Revert "test(inspector): run console resume outside recorder cleanup" #43886 at fb783bb, using the same extension revision, but in the default project, not default-reuse. The database has no VS Code results, and I did not establish an independent failure of the exact default-reuse test. Under the guide's evidence rule, this remains uncertain rather than a confirmed flake. A base/head rerun of this exact project would resolve that gap.

CI failures checked by the Playwright bot - agent run

Triaged by the Playwright bot - agent run

@github-actions

github-actions Bot commented Nov 3, 2026

Copy link
Copy Markdown
Contributor

🟢 Neither failure looks caused by this PR

Hi, I'm the Playwright bot. I checked both failures for a355853 against the diff and independent CI runs.

Details

The PR restores skips in after hooks and removes two tests for the reverted restriction. Neither failure comes from an after-hook skip.

Pre-existing flake / infra

  • Firefox: [firefox] › mcp/cli-session.spec.ts:148 › session stops when socket directory in TMPDIR disappears fails because the daemon remains alive after the five-second wait. The refreshed test-results database shows the exact test and Ubuntu Firefox configuration failing with the same assertion on main at 47ba108, before this PR. Excluding this PR, it failed in 5 of 193 runs and passed in 188. This is a confirmed pre-existing flake.

  • VS Code: [default-reuse] › run-tests.spec.ts:1041 › should produce output twice fails because an extra GitCommitInfo: timeout ... git fetch origin db6633be... line appears in the output; the inner test passes. The same project and test fail with the same extra line on independent Revert "feat(locator): add a context-wide visibility default" #43888 at c96d4f9, in this job. That commit still has _allowSkips = false, so this PR's change cannot explain that failure. Git metadata collection runs during setup, before the changed after-hook code. This is an existing Git-fetch timeout affecting the output assertion, not a regression from this PR. VS Code results are absent from the database, so this comparison uses job logs.

Triaged by the Playwright bot - agent run

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants