[ty] propagate classmethod-ness through decorators returning Callables (#21958)

Fixes https://github.com/astral-sh/ty/issues/1787

## Summary

Allow method decorators returning Callables to presumptively propagate
"classmethod-ness" in the same way that they already presumptively
propagate "function-like-ness". We can't actually be sure that this is
the case, based on the decorator's annotations, but (along with other
type checkers) we heuristically assume it to be the case for decorators
applied via decorator syntax.

## Test Plan

Added mdtest.
This commit is contained in:
Carl Meyer
2025-12-16 09:16:40 -08:00
committed by GitHub
parent 7f7485d608
commit eeaaa8e9fe
6 changed files with 75 additions and 24 deletions

View File

@@ -113,9 +113,7 @@ intention that it shouldn't influence the method's descriptor behavior. For exam
```py
from typing import Callable
# TODO: this could use a generic signature, but we don't support
# `ParamSpec` and solving of typevars inside `Callable` types yet.
def memoize(f: Callable[[C1, int], str]) -> Callable[[C1, int], str]:
def memoize[**P, R](f: Callable[P, R]) -> Callable[P, R]:
raise NotImplementedError
class C1:
@@ -259,4 +257,37 @@ reveal_type(MyClass().method) # revealed: (...) -> int
reveal_type(MyClass().method.__name__) # revealed: str
```
## classmethods passed through Callable-returning decorators
The behavior described above is also applied to classmethods. If a method is decorated with
`@classmethod`, and also with another decorator which returns a Callable type, we make the
assumption that the decorator returns a callable which still has the classmethod descriptor
behavior.
```py
from typing import Callable
def callable_identity[**P, R](func: Callable[P, R]) -> Callable[P, R]:
return func
class C:
@callable_identity
@classmethod
def f1(cls, x: int) -> str:
return "a"
@classmethod
@callable_identity
def f2(cls, x: int) -> str:
return "a"
# error: [too-many-positional-arguments]
# error: [invalid-argument-type]
C.f1(C, 1)
C.f1(1)
C().f1(1)
C.f2(1)
C().f2(1)
```
[`tensorbase`]: https://github.com/pytorch/pytorch/blob/f3913ea641d871f04fa2b6588a77f63efeeb9f10/torch/_tensor.py#L1084-L1092

View File

@@ -444,20 +444,18 @@ When a `@classmethod` is additionally decorated with another decorator, it is st
class method:
```py
from __future__ import annotations
def does_nothing[T](f: T) -> T:
return f
class C:
@classmethod
@does_nothing
def f1(cls: type[C], x: int) -> str:
def f1(cls, x: int) -> str:
return "a"
@does_nothing
@classmethod
def f2(cls: type[C], x: int) -> str:
def f2(cls, x: int) -> str:
return "a"
reveal_type(C.f1(1)) # revealed: str