Release 9.0.0

[github-actions]

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-9.0.0.
• Release version: 9.0.0.

Or execute on the command line:

gh workflow run deploy.yml -r release-9.0.0 -f version=9.0.0

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

[bluetech]

I will edit the changes later, but any comments welcome in the meantime.

The RTD build has some error in subtests.rst:

/home/docs/checkouts/readthedocs.org/user_builds/pytest/checkouts/13878/doc/en/how-to/subtests.rst:29: WARNING: Lexing literal_block '$ pytest -q test_subtest.py\nuuuuuF                                                               [100%]\n================================= FAILURES =================================\n_______________________ test [custom message] (i=1) ________________________\n\nsubtests = <_pytest.subtests.Subtests object at 0xdeadbeef0001>\n\n    def test(subtests):\n        for i in range(5):\n            with subtests.test(msg="custom message", i=i):\n>               assert i % 2 == 0\nE               assert (1 % 2) == 0\n\ntest_subtest.py:6: AssertionError\n_______________________ test [custom message] (i=3) ________________________\n\nsubtests = <_pytest.subtests.Subtests object at 0xdeadbeef0001>\n\n    def test(subtests):\n        for i in range(5):\n            with subtests.test(msg="custom message", i=i):\n>               assert i % 2 == 0\nE               assert (3 % 2) == 0\n\ntest_subtest.py:6: AssertionError\n___________________________________ test ___________________________________\ncontains 2 failed subtests\n========================= short test summary info ==========================\nSUBFAILED[custom message] (i=1) test_subtest.py::test - assert (1 % 2) == 0\nSUBFAILED[custom message] (i=3) test_subtest.py::test - assert (3 % 2) == 0\nFAILED test_subtest.py::test - contains 2 failed subtests\n3 failed, 3 subtests passed in 0.12s' as "pytest" resulted in an error at token: 'u'. Retrying in relaxed mode. [misc.highlighting_failure]

[nicoddemus]

The RTD build has some error in subtests.rst:

Seem related to the new u character that is used as short letter for a failed subtest:

uuuuuF                                                               [100%]

We might need to update the pytest highlighter, perhaps @asottile can help?

Otherwise we can change from .. code-block:: pytest to .. code-block:: console and lose the syntax highlighting.

[bluetech]

I addressed @nicoddemus's comments and edited the changelog.

The other parts of the diff have some problems I need to check.

[asottile]

I can take a look at the highlighter later today probably

[bluetech]

I tried to fix regendoc to handle nested directives but my simplistic attempt (fiddling with the dedent function) fails on the full pytest docs. regendoc works on the textual level and treats a consecutive sequence of lines with a non-zero indentation (modulo empty lines) to be a "block". That's doesn't handle nesting. Need to think of another approach.

[asottile]

latest pygments-pytest has ~partial support for subtests. at least for now I didn't bother handling all the possible highlighting but just the basic ones to unblock this

[kieran-ryan]

Suggested change

	pytest 8.5.0.dev291+gadb3658f0.d20251105
	pytest 9.0.0

[bluetech]

Hmm I think we have a SETUPTOOLS_SCM_PRETEND_VERSION thing that should have taken care of it. I will check.

[bluetech]

I'm not sure why it doesn't work, I can reproduce that SETUPTOOLS_SCM_PRETEND_VERSION_FOR_PYTEST=9.0.0 tox -e regendoc doesn't make it 9.0.0. I changed it manually for now, don't want to go into another rabbit hole :)

[webknjaz]

@bluetech details: https://github.com/pytest-dev/pytest/pull/13928/files#r2515081435

[bluetech]

@asottile Thanks, updated the bound to pygments-pytest>=2.5.0.

[bluetech]

Released to pypi. The "create github release" failed, but too late for here, so I'll merge the PR and handle the github release manually this once...

[Pierre-Sassoulas]

Failed because git is not installed in the release image (probably the same reason than setuptools_scm fails ?)

[bluetech]

I think it rather needs a clone of the repo (I thought it didn't if you specify a pre-existing tag but guess not).

[bluetech]

I'll merge the PR and handle the github release manually this once...

Done.