I think we used to have this but it was removed. Or at least, it's usually not a problem since tests should clean up after themselves. If they don't then ignoring it just hides the problem and lets it grow.

The /test/old/testdir/X* case is an exception since those tests are owned by upstream.

If they don't then ignoring it just hides the problem and lets it grow.

Yeah, but I would argue that it causes more inconvenience and pollutes the git status output, even when you're not working with these tests. Such "broken" tests can easily be merged, since there is no CI to ensure proper cleanup. To me, it still makes sense to gitignore generated files that are never meant to be staged, because the test runner can be killed or fail to delete them for other reasons.

Anyway, if they aren't supposed to remain, I'll submit a fix to ensure proper test cleanup.

are you aware that you can run a single test file? https://github.com/neovim/neovim/blob/master/test/README.md#filter-by-file

are you aware that you can run a single test file?

ofc, but I still run all tests time to time to ensure that other parts, which I might not notice are affected, are still working.

Also, at least two of these tests are not properly cleaning up because they rely on vim.uv.sleep to wait for nvim to exit. On my machine, it takes longer to close, so the logs are being written after after_each executes. Ignoring them with .gitignore is not really a solution, the real fix would be to handle subprocesses exit properly and wait before cleanup, but I don't think this is a high-priority problem right now.

Maybe the compromise is to temporarily ignore them. Moving these files into a subdir could also make things cleaner.