Specific ruff file ignores ignore everything

[henryiii]

The following file:

def main(x):
    a = 1

Shows two errors.

$ ruff check --isolated --extend-select ARG tmp.py
tmp.py:2:10: ARG001 Unused function argument: `x`
tmp.py:3:5: F841 [*] Local variable `a` is assigned to but never used
Found 2 errors.
[*] 1 potentially fixable with the --fix option.

If I add this to the top of the file:

# ruff: noqa: F821

Then both errors disappear, instead of just one. In general, it seems that this is just a blanket noqa. (Also, this should not be too picky - ruff: noqa F821 is an easy mistake to make! I think I'd require a # to put a comment on the same line, if it was my choice).

See pybind/pybind11#4483 (comment).

[charliermarsh]

Do you know what version this is on?

[charliermarsh]

In #2978, we added support for directives like # ruff: noqa: F821 which ignore a specific error.

(Note that Flake8 treats # flake8: noqa: F821 as a blanket ignore -- as in, with Flake8, that directive will ignore all errors like you're seeing above.)

[henryiii]

Locally, my version is a bit older due to being from homebrew - 0.0.245. In pre-commit in the original issue, it should have been the latest version, but I can check - ahh, no, it's 0.0.247. I knew that was recent, but not quite how recent. It would be really helpful if # ruff: ignore<stuff> was an error unless <stuff> was recognized and supported or started with a #.

I'll check and report back later tonight!

[charliermarsh]

So as of that PR, we do throw warnings -- e.g., if you do # ruff: noqa F401:

warning: Unexpected suffix on `noqa` directive: "# ruff: noqa F401"

[charliermarsh]

(I think that feature shipped in 0.0.248.)

[henryiii]

It oddly prints the warning three times:

warning: Unexpected suffix on `noqa` directive: "# ruff: noqa SIM201 SIM300 SIM202"
warning: Unexpected suffix on `noqa` directive: "# ruff: noqa SIM201 SIM300 SIM202"
warning: Unexpected suffix on `noqa` directive: "# ruff: noqa SIM201 SIM300 SIM202"

Fixed 19 errors:
- tests/test_enum.py:
    10 × SIM201 (negate-equal-op)
     7 × SIM300 (yoda-conditions)
     2 × SIM202 (negate-not-equal-op)

(And I'd be quite fine to have some sort of way to make all warnings errors).

But it does work correctly, and warns as expected, in 0.0.249 (and therefore I'd assume 0.0.251)! Sorry for the noise, I should have tried an update first.