Skip to content

Conversation

@sanchit122006
Copy link

@sanchit122006 sanchit122006 commented Dec 18, 2025

This pull request makes the test subprocess helper a bit smarter on Windows to avoid spurious timeouts.

Added subprocess section in subprocess_run_for_testing function

image

Modified subprocess section

image

This change updates subprocess_run_for_testing so that, on Windows, it runs test subprocesses with the CREATE_NO_WINDOW flag. This prevents console windows from being created for every subprocess, which can contribute to the hanging and timeout behavior seen in the Windows test runs. The new behavior is guarded by sys.platform == "win32" and merges CREATE_NO_WINDOW with any existing creationflags, so non‑Windows platforms and callers that already pass creationflags keep their current behavior.

PR checklist

@timhoffm
Copy link
Member

@sanchit122006 are you posting unreviewed tooling / AI output here? E.g.

  • This has nothing to do with subprocesses and windows:

    Fix macOS CI: remove redundant brew update and add coverage file check

  • There are a lot of formatting / style changes, which totally obscure whatever logic might be here
  • Why are there code/docstring screenshots without explanation?

This is not acceptable. I put the PR in draft state for now. Please cleanup and make sure all content (code and PR message are presented in a clean way). When you think the content is suitably consolidated so that it's worth putting valuable core developer time into, you can remove the draft status.

@timhoffm timhoffm marked this pull request as draft December 18, 2025 13:15
@sanchit122006
Copy link
Author

sanchit122006 commented Dec 18, 2025

@timhoffm Can you guide me to solve cl check's issue ,
macos cl, whitespaces and iteration's commits is for solving command line check failing issue of this PR ,
should i not commit the changes i made in action files to solve command line issue?
could checks passed by only saving the file and not commiting?

@timhoffm
Copy link
Member

macos cl, whitespaces and iteration's commits is for solving command line check failing issue of this PR , should i not commit the changes i made in action files to solve command line issue?

What exactly is "command line check failing issue"? Are these failures in your local setup, e.g. pre-commit?

Generally, a PR should only contain changes that are related to the PR. If for example, I very stongly assume the brew update change is independent of this PR. If you think a such a change is reaonable (or even prerequisite but independently meaningful) , please open a separate PR. It's much simpler to handle logically independent change separately.

@sanchit122006
Copy link
Author

sanchit122006 commented Dec 18, 2025

@timhoffm i am talking about these checks, can you guide me should i make new branch and PR on it, because 1 test is failing yet before has 2 failing checks then i edited and commit to fix macos check
image

@timhoffm
Copy link
Member

Unfortunately, this is another unrelated flaky behavior: #30882. You can ignore this for the scope of this PR.

Likewise the osx issue is unrelated to the PR.

e.g. both have also occured in #30600.

@sanchit122006
Copy link
Author

sanchit122006 commented Dec 19, 2025

@timhoffm Refs #30884 Check this out dear i made clean PR with clean changes
Hope this help !!

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.

[Bug]: Random timeout failures in CI

2 participants