F401 false positives when match statements are in the file

[kendangclubhouse]

Im seeing that ruff is not finding F401's in files that have match statements.

Code snippet

import datetime

x = 5
match x:
    case _:
        print("hello")

You should expect that import datetime should cause a F401 but when runing ruff /path/to/file --select F401 nothing is returned. When I remove the match statement, ruff will return the F401 issue.

❯ ruff ~ruff_bug.py --select F401
ruff_bug.py:1:8: F401 `datetime` imported but unused
  |
1 | import datetime
  |        ^^^^^^^^ F401
  |
  = help: Remove unused import: `datetime

Im running ruff 0.0.225

[charliermarsh]

I believe this is due to our lack of support for match statements. It's the last big language feature that we don't yet support (#282). I assume that if you --select E999, you'll see a syntax error? That's probably blocking the F401 from being picked up.

[kendangclubhouse]

I find it odd that I cannot ignore the E999. Does E999 have to block all other errors in the file?

[charliermarsh]

The issue is that if we can't parse the code, then we can't get a valid AST, and so we can't enforce a large portion of the rule set. It's a limitation of the parser right now: (1) it can't parse match statements, and (2) it can't recover from errors gracefully. (Even if it could, you would still end up with false positives or negative -- for example, if an import was used only in a match statement, we'd have no way of knowing.)

[kendangclubhouse]

I see, that make a lot of sense. Thank you so much for explaining. 🙂

[charliermarsh]

No prob! I know it’s not a satisfying answer, but hopefully it makes some sense. The real fix here is to get match statements working so that it’s a non-issue!

[charliermarsh]

Fixed as of v0.0.250.