Support for validate_call with generics

[caniko]

Initial Checks

• I confirm that I'm using Pydantic V2

Description

validate_call should catch wrong types being passed to any function/method; however, it breaks when used with parametrized generic methods.

A similar issue occurs when generic functions have incorrect return type.

Example Code

from typing import TypeVar, Generic

from pydantic import BaseModel, validate_call

T = TypeVar("T")


class MyGModel(BaseModel, Generic[T]):
    @validate_call
    def my_generic_method(self, arg: T) -> None:
        return


# Should fail because 1 is not str
MyGModel[str]().my_generic_method(1)


@validate_call(validate_return=True)
def my_func(arg: T) -> T:
    return 1


# Should fail because "5" is str, but 1 is int
my_func("5")

Python, Pydantic & OS Version

pydantic version: 2.4.2
        pydantic-core version: 2.10.1
          pydantic-core build: profile=release pgo=true
                 install path: /home/can/.cache/pypoetry/virtualenvs/chainspect-uudSkfMt-py3.11/lib/python3.11/site-packages/pydantic
               python version: 3.11.4 (main, Jun 19 2023, 07:09:29) [GCC 13.1.1 20230429]
                     platform: Linux-6.5.6-zen2-1-zen-x86_64-with-glibc2.38
             related packages: mypy-1.5.1 typing_extensions-4.8.0

[sydney-runkle]

@caniko,

Thanks for reporting this. It does seem to be a bug 🐛.

[sydney-runkle]

Going to move this to a feature-request instead - though I think it's one that should be relatively high on our list 😄.

[AmailP]

Hi @sydney-runkle, since the #7928 has been closed, would it make sense to add this issue to a milestone, since it seems to be your new approach to issues?

[sydney-runkle]

@AmailP,

This isn't a super high priority issue at the moment, so I don't think we'll add it to the 2.7 milestone. If there's more community support, or if someone wants to take this on, we can certainly move up the timeline!

[sydney-runkle]

A good example of this lack of support: #9536

[sydney-runkle]

I'm going to add this to v2.9 - this is something I'd like to work on, though it will take a back seat compared to performance concerns in the 2.9 sprint.