False positive A003 when overriding threading.Event.set

[jamesbraza]

from threading import Event

class MyEvent(Event):
    def set(self) -> None:
        super().set()
        print("My override")

Running ruff==0.0.279:

> ruff --isolated --select="A003" a.py
a.py:5:9: A003 Class attribute `set` is shadowing a Python builtin

This seems like a false positive, as one has to use the word set to override the Event parent class's method.

Should A003 be raised on method overrides?

[charliermarsh]

Probably not... but somewhat low-priority for me, since we can't resolve the superclasses anyway and it's not a hugely popular rule. The most lightweight thing I could see us doing for now is ignoring methods annotated with @overrides.

[jamesbraza]

Just out of curiosity, why can't ruff resolve superclasses? I believe other linters like pylint have this capability

[charliermarsh]

Ruff doesn't resolve symbols across files right now, though it will in the future (I've been working on it a bit but there's no defined timeline for it).

pylint does resolve symbols across files and does more static analysis than other linters, though not all linters function that way -- Flake8 (as an alternative example) couldn't support this.

[charliermarsh]

I will special-case some stuff in the standard library for now.