Skip to content

Conversation

@github-actions
Copy link
Contributor

Created by the prepare release pr workflow.

Once all builds pass and it has been approved by one or more maintainers, start the deploy workflow, using these parameters:

  • Use workflow from: release-9.0.1.
  • Release version: 9.0.1.

Or execute on the command line:

gh workflow run deploy.yml -r release-9.0.1 -f version=9.0.1

After the workflow has been approved by a core maintainer, the package will be uploaded to PyPI automatically.

@nicoddemus nicoddemus marked this pull request as ready for review November 11, 2025 13:07
$ pytest --version
pytest 9.0.0
pytest 9.0.1.dev17+g0fa11ae3f.d20251111
Copy link
Member

@nicoddemus nicoddemus Nov 11, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@RonnyPfannschmidt we use the SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST trick to force the version (#8134), do you know if this is still supposed to be working?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can this be due to https://github.com/pytest-dev/pytest/pull/13841/files#diff-ef2cef9f88b4fe09ca3082140e67f5ad34fb65fb6e228f119d3812261ae51449R98 being built in a separate environment where set_env on [testenv:regen] doesn't apply?

Try setting it in a [pkgenv] section as follows:
https://github.com/ansible/awx-plugins/blob/5b5f2eb1baa62d8d12805cf8da7c076aee9d1b3e/tox.ini#L142-L151

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The diff for tox c -e .pkg before and after 68016f0 is

--- pre-cutoff-config-.pkg.log	2025-11-11 18:34:47.680413742 +0100
+++ cutoff-config-.pkg.log	2025-11-11 18:34:55.967007776 +0100
@@ -2,7 +2,7 @@
 type = Pep517VirtualEnvPackager
 set_env =
   PIP_DISABLE_PIP_VERSION_CHECK=1
-  PYTHONHASHSEED=3427332812
+  PYTHONHASHSEED=1574684877
   PYTHONIOENCODING=utf-8
   PYTHONWARNDEFAULTENCODING=1
   _PYTEST_TOX_DEFAULT_POSARGS=   
@@ -83,4 +83,4 @@
 config_settings_build_editable = None
 fresh_subprocess = True
 deps = 
-# !!! unused: commands, description, extras
+# !!! unused: commands, description, extras, package, wheel_build_env

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

And it does contain

pass_env =
  SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST

in both cases.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The last release that didn't have a generated version in the patch was v8.4.2 @ #13696. In the CI log, it's visible that it also used a Git-derived on the tox level — https://github.com/pytest-dev/pytest/actions/runs/17435825368/job/49505262303#step:5:93:

regen: install_package> python -I -m pip install --force-reinstall --no-deps /home/runner/work/pytest/pytest/.tox/.tmp/package/3/pytest-8.4.2.dev21+g89905381a.d20250903.tar.gz

However the regendoc invocation used the requested version. It was using tox v4.29.0 and newer runs use tox v4.32.0. The rest of the dep tree is the same.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I confirmed that using tox==4.29.0 does not have this problem.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

tox==4.31 is fine too

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Apparently, I can't reproduce this with SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST=9.0.1 python -Im tox r -e regen even with tox==4.32. So it's the SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST=9.0.1 tox r -e prepare-release-pr --notest invocation that might not be passing the var into regen.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@nicoddemus I triggered the workflow to test #13933 and it apparently updated this PR. I didn't anticipate this but on the bright side — we now do know that it works. I'm going to merge the fix in, so it's absorbed properly and delete the experimental branch release-experiment/pr-13933 too.

@webknjaz
Copy link
Member

@nicoddemus once #13935 is in, we should probably trigger pre-release again so it originates from a proper branch.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@webknjaz webknjaz dismissed nicoddemus’s stale review November 11, 2025 20:20

This concern has been resolved by #13935.

@webknjaz
Copy link
Member

I've triggered https://github.com/pytest-dev/pytest/actions/runs/19277570192/job/55121005018 — it will update this PR shortly.

@webknjaz
Copy link
Member

Looks like the release script doesn't handle existing PRs gracefully: https://github.com/pytest-dev/pytest/actions/runs/19277639565/job/55121223740#step:5:4792. It does push the branch but still attempts to create a PR which fails. Instead, it should try updating the PR description, if applicable or just do nothing..

@nicoddemus
Copy link
Member

Thanks a lot @webknjaz for all the digging and ultimately finding a solution and merging it!

Instead, it should try updating the PR description, if applicable or just do nothing..

I agree, but for now I will just close this PR and start from scratch.

@nicoddemus nicoddemus closed this Nov 11, 2025
@nicoddemus nicoddemus deleted the release-9.0.1 branch November 11, 2025 22:29
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.

4 participants