revert: remove scheduler overhead tolerance#153
Conversation
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 641321156a
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| 1000 * tasksFinished + | ||
| 1000 + | ||
| schedulerOverheadMargin, | ||
| tasksFinishedTimeout + 1000 * tasksFinished + 1000, |
There was a problem hiding this comment.
This test now fails whenever the two timeout-based waits in destroyWorkerNode() overshoot by even a few milliseconds: waitWorkerNodeEvents(..., tasksFinishedTimeout, false) and the subsequent kill-message timeout both use setTimeout, and the CI workflow runs deno task test:coverage across Windows, macOS, Ubuntu, and multiple Deno versions. On a busy runner, an elapsed time like 2001ms is expected scheduler delay rather than a destroy() regression, so removing the small margin reintroduces a flaky CI failure.
Useful? React with 👍 / 👎.
1 participant
Reverts the Hermes-introduced test timing tolerance from PR #152, restoring the previous assertion exactly.