Allow typing.Final for mutable-class-default annotations (RUF012)

[charliermarsh]

Summary

See: #5243.

[github-actions]

PR Check Results

Ecosystem

✅ ecosystem check detected no changes.

Benchmark

Linux

group                                      main                                   pr
-----                                      ----                                   --
formatter/large/dataset.py                 1.00      6.9±0.01ms     5.9 MB/sec    1.00      6.9±0.04ms     5.9 MB/sec
formatter/numpy/ctypeslib.py               1.00   1422.9±4.62µs    11.7 MB/sec    1.01   1430.6±3.04µs    11.6 MB/sec
formatter/numpy/globals.py                 1.00    133.2±0.22µs    22.2 MB/sec    1.00    133.4±0.65µs    22.1 MB/sec
formatter/pydantic/types.py                1.00      2.9±0.01ms     8.9 MB/sec    1.00      2.9±0.01ms     8.9 MB/sec
linter/all-rules/large/dataset.py          1.00     13.6±0.06ms     3.0 MB/sec    1.01     13.7±0.06ms     3.0 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      3.5±0.01ms     4.8 MB/sec    1.01      3.5±0.00ms     4.7 MB/sec
linter/all-rules/numpy/globals.py          1.00    362.8±5.82µs     8.1 MB/sec    1.00    362.5±1.34µs     8.1 MB/sec
linter/all-rules/pydantic/types.py         1.00      6.1±0.02ms     4.2 MB/sec    1.00      6.1±0.03ms     4.2 MB/sec
linter/default-rules/large/dataset.py      1.00      7.1±0.02ms     5.8 MB/sec    1.00      7.1±0.07ms     5.7 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00  1483.4±12.75µs    11.2 MB/sec    1.01   1493.9±2.15µs    11.1 MB/sec
linter/default-rules/numpy/globals.py      1.00    158.1±0.18µs    18.7 MB/sec    1.00    158.8±0.13µs    18.6 MB/sec
linter/default-rules/pydantic/types.py     1.00      3.2±0.01ms     7.9 MB/sec    1.00      3.2±0.01ms     7.9 MB/sec

Windows

group                                      main                                   pr
-----                                      ----                                   --
formatter/large/dataset.py                 1.08      8.3±0.07ms     4.9 MB/sec    1.00      7.7±0.06ms     5.3 MB/sec
formatter/numpy/ctypeslib.py               1.06  1730.3±19.65µs     9.6 MB/sec    1.00  1636.7±19.16µs    10.2 MB/sec
formatter/numpy/globals.py                 1.03    164.0±2.47µs    18.0 MB/sec    1.00   159.7±11.17µs    18.5 MB/sec
formatter/pydantic/types.py                1.07      3.5±0.05ms     7.4 MB/sec    1.00      3.2±0.03ms     7.9 MB/sec
linter/all-rules/large/dataset.py          1.00     15.0±0.09ms     2.7 MB/sec    1.00     14.9±0.08ms     2.7 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.01      4.0±0.05ms     4.2 MB/sec    1.00      4.0±0.03ms     4.2 MB/sec
linter/all-rules/numpy/globals.py          1.00    484.6±5.36µs     6.1 MB/sec    1.00    483.9±8.98µs     6.1 MB/sec
linter/all-rules/pydantic/types.py         1.01      6.7±0.30ms     3.8 MB/sec    1.00      6.6±0.05ms     3.8 MB/sec
linter/default-rules/large/dataset.py      1.02      7.9±0.07ms     5.1 MB/sec    1.00      7.7±0.05ms     5.3 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.01  1680.3±13.55µs     9.9 MB/sec    1.00  1658.7±16.21µs    10.0 MB/sec
linter/default-rules/numpy/globals.py      1.01    192.2±2.98µs    15.3 MB/sec    1.00    191.1±3.75µs    15.4 MB/sec
linter/default-rules/pydantic/types.py     1.02      3.6±0.07ms     7.2 MB/sec    1.00      3.5±0.03ms     7.3 MB/sec

[andersk]

Final doesn’t mean immutable; Final[list[int]] is just as mutable as list[int]. So I don’t see why it should get a special exemption here.

from typing import Final

class Foo:
    var: Final[list[int]] = []

a = Foo()
a.var.extend([1, 2, 3])
b = Foo()
print(b.var)  # [1, 2, 3]

(mypy playground)

var: Final[list[int]] protects against reassignment like a.var = [1, 2, 3], but that has nothing to do with list’s mutability: var: Final[int] and a.var = 2 is equally problematic.

[charliermarsh]

Do you think ClassVar should have a special exemption here?

[charliermarsh]

The reason I'm asking is that PEP 591 says:

Type checkers should infer a final attribute that is initialized in a class body as being a class variable. Variables should not be annotated with both ClassVar and Final.

That is: if a variable is annotated with Final and initialized in the class body, it implies ClassVar, which we are exempting.

[andersk]

Good question. By my above logic, ClassVar should not be exempted, but maybe the idea is that we only want to protect against the simultaneous possibility of both mutation and reassignment, in which case ClassVar and Final are fine? I withdraw this criticism for now.