pre-commit check for xfail in unit tests

[frankcash]

will resolve #28768

Adds regexp check for @pytest.mark.xfail in files ^tests/.*\.py$

---

^ Add meaningful description above

Read the Pull Request Guidelines for more information.
In case of fundamental code changes, an Airflow Improvement Proposal (AIP) is needed.
In case of a new dependency, check compliance with the ASF 3rd Party License Policy.
In case of backwards incompatible changes please leave a note in a newsfragment file, named {pr_number}.significant.rst or {issue_number}.significant.rst, in newsfragments.

[eladkal]

We still have 1 test that needs to be excluded #23539 as we didnt finish removing xfail everywhere

[frankcash]

@eladkal we could also wait to merge this until #28724 is merged

[Taragolis]

We also have one xfail marked test which not listed in #23539

airflow/tests/always/test_project_structure.py

Lines 377 to 379 in 3eee33a

	@pytest.mark.xfail(reason="We did not reach full coverage yet")
	def test_missing_assets(self):
	super().test_missing_assets()

One of suggestion is remove this kind of tests, see: #28459 (comment)

[eladkal]

@frankcash I suggest to add exclusion for the 1-2 remaining tests
cleaning it after tests are refactored is simple

[frankcash]

@eladkal does this look like adding those files specifically to the exclude file list? exclude: ^airflow/|^docs/. or extend the regex?

[frankcash]

@eladkal ✅

[ashb]

This is a -0.99 from me. XFail exists for a reason, we can discourage it's use without needing to remove it.

See #28840 for where I think an xfail makes sense.

[potiuk]

This is a -0.99 from me. XFail exists for a reason, we can discourage it's use without needing to remove it.

Yep. I think there are reasons why you might want to have xfail. Some of the other fixes xfails in #23539 were simply mis-used - their reason was "the tests were flaky" but using xfail for the reason of "we know it should fail and we can fix it later" is quite ok.

[potiuk]

BTW. it does not mean that the encouragement of getting rid of those xfals is bad idea (if done correctly).

[ashb]

Oh yeah I'm cool with "don't use XFail for flaky tests" etc.