Use pytest 8 in CI

[mattwthompson]

• Tag issue being addressed
• Add tests
• Update docstrings/documentation, if applicable
• Lint codebase
• Update changelog

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 94.08%. Comparing base (9b1ab19) to head (05c2fc2).
Report is 1 commits behind head on main.

Additional details and impacted files

[mattwthompson]

Ha, I've gotten this error before but in a position where it didn't need to be fixed

=========================== short test summary info ============================
FAILED openff/toolkit/_tests/test_forcefield.py::TestForceField::test_load_bad_filelike_object - TypeError: argument of type 'NoneType' is not iterable
FAILED openff/toolkit/_tests/test_forcefield.py::TestForceField::test_do_not_load_in_child_dir - TypeError: argument of type 'NoneType' is not iterable
FAILED openff/toolkit/_tests/test_forcefield.py::TestForceField::test_load_bad_bytes - TypeError: argument of type 'NoneType' is not iterable
FAILED openff/toolkit/_tests/test_forcefield.py::TestForceField::test_load_bad_string - TypeError: argument of type 'NoneType' is not iterable

[mattwthompson]

Splitting off some more discrete ideas which might be nice to have in the installed package before this works nicely, starting with

• Skip some tests if unpackaged files not present, make some files path-y-er #1825
• Allow writing molecules to paths #1826

[j-wags]

Thanks, this was clearly a ton of work all over the codebase and I really appreciate you hunting it all down. However in the current state I can't tell what's a temporary patch to get CI passing vs. what would actually be committed. Now that I see the green CI on this PR, could you modify it to be in its final state and update the PR text with any major subsequent steps that you'd plan to take after the review (otherwise I can't tell what's a temporary patch vs. a genuine oversight).

[j-wags]

(blocking) Could you expand on the reasoning for not putting these in conda_oe.yaml? Especially with version pins this seems like something that may confuse us in the future. If these are outside of that file intentionally, it'd be good to leave a comment explaining why.

[mattwthompson]

I wanted to drop that file but instead I'm just reverting all of the "conda" tests

[j-wags]

(blocking) This line seems essential to ensuring that the last release is being tested against its own tests, rather than the tests currently on main. Could you clarify?

[mattwthompson]

As above

[j-wags]

Awesome, thanks for all the hard work and for iterating on this @mattwthompson. Please update the releasenotes before you merge!