Running black on this to solve the line length issues (while astral-sh/ruff#6936 won't be implemented in the near future) produced some interesting results: black.patch

Not entirely mergeable either since this also reformats the style of parameter indentation in function calls.

What is wrong with

reformats the style of parameter indentation in function calls.

as long as it is automatic and consistent with other Python codebases?

It would be a new inconsistency introduced to our own codebase. We have to decide whether we want that, to which I’m neutral.

On #317, which the unittest changes are blocking, I said:

That is still a draft, no? Do you want to me to replace your PR's unittest file with the black-formatted one I provided and then merge it?

If you don't object, I'll do that when I get home in 8 hours today.

On second thought, I think I should implement the unittest dropping on the refactoring PR instead, since it's more topic-related and our branches conflict quite a bit (plus some of the changes here were already made in the refactoring branch).

...that turned out to be an even worse idea since this PR does replace the asserts. I'll just do the black thing then.

Do we need to do the setUp() and tearDown() for every little small test, as it is in with the run_around_tests() fixture rn? I feel like it'd do if these functions only run once every session.

Let’s put this PR on hold. It is still draft. We can merge #317 first.

• Refactor tests and metadata #317

Sorry, I didn't realize that you had reversed the changes removing unittest from unittests but also couldn't make a proposal as the relevant lines are outside the diff range. Treat my commit as you wish.

doesn't fix running with pytest

Local testing I do:
% python3 -m venv .venv
% source .venv/bin/activate
% python -m pip install --editable ".[dev]"
% pytest

python cpplint_clitest.py and python cpplint_unittest.py also work in an activated venv.