Issue with constrained type var and Protocol

[Redoubts]

Bug Report

Seems like a regression from 11.2, consider the following

import gzip, shutil
from pathlib import Path


with open("/dev/null") as src, gzip.open("/dev/null", "wt") as dest:
    reveal_type(src)
    reveal_type(dest)
    shutil.copyfileobj(src, dest)


with Path("/dev/null").open() as src2, gzip.open("/dev/null", "wt") as dest2:
    reveal_type(src2)
    reveal_type(dest2)
    shutil.copyfileobj(src2, dest2)

I see errors even though this should be fine?

% mypy x.py     
x.py:6: note: Revealed type is "io.TextIOWrapper[io._WrappedBuffer]"
x.py:7: note: Revealed type is "typing.TextIO"
x.py:8: error: Cannot infer type argument 1 of "copyfileobj"  [misc]
x.py:12: note: Revealed type is "io.TextIOWrapper[io._WrappedBuffer]"
x.py:13: note: Revealed type is "typing.TextIO"
x.py:14: error: Cannot infer type argument 1 of "copyfileobj"  [misc]
Found 2 errors in 1 file (checked 1 source file)

no errors in 1.11.2:

x.py:6: note: Revealed type is "io.TextIOWrapper"
x.py:7: note: Revealed type is "typing.TextIO"
x.py:12: note: Revealed type is "io.TextIOWrapper"
x.py:13: note: Revealed type is "typing.TextIO"
Success: no issues found in 1 source file

Your Environment

• Mypy version used: 1.13
• Mypy command-line flags: mypy
• Mypy configuration options from mypy.ini (and other config files): none
• Python version used: Python 3.11.10

[brianschubert]

Looks potentially related to python/typeshed#12286 (which landed in v1.12.0 with #17586)

[hauntsaninja]

mypy_primer -p test.py --bisect --debug --old 'v1.11.0' confirms it's the sync PR brianschubert mentions

[JukkaL]

Here's an attempt to create a simplified, self-contained reproducer:

from typing import AnyStr, Protocol, TypeVar, Generic, overload

class Buffer(Protocol):
    def __buffer__(self, flags: int, /) -> memoryview: ...

class IO(Generic[AnyStr]):
    @overload
    def write(self: IO[bytes], s: Buffer, /) -> int: ...
    @overload
    def write(self, s: AnyStr, /) -> int: ...
    def write(self, s):
        return 0

class TextIO(IO[str]):
    pass

_T_contra = TypeVar("_T_contra", contravariant=True)

class SupportsWrite(Protocol[_T_contra]):
    def write(self, s: _T_contra, /) -> object: ...

def foo(
    f: SupportsWrite[AnyStr],
) -> None: ...

def bar(d: TextIO) -> None:
    foo(d)  # Value of type variable "AnyStr" of "foo" cannot be "Buffer"

Interestingly, this fails on 1.11 as well, and doesn't depend on the changes in #17586.

I looked at this a bit, it seems that constraints inference from protocol generates invalid constraints (infer_constraints_from_protocol_members). The root of the problem seems to be in mypy.subtypes.find_member.

@ilevkivskyi I remember that you mentioned that you've looked into inconsistencies between find_member and analyze_member_access. This may be an example where they deviate.

[ilevkivskyi]

I remember that you mentioned that you've looked into inconsistencies between find_member and analyze_member_access. This may be an example where they deviate.

Yes, and it may not be trivial to fix. Ultimately, we need to refactor this, so that everything uses analyze_member_access (e.g. also sub-classing overrides checks, although they are currently closer in semantics/implementation). As you know, this is a big project, and I didn't yet have courage to start it :-)

[JukkaL]

It may still be possible to fix this specific regression by a more ad-hoc fix, by adding some functionality to find_member, even if we don't fully migrate to analyze_member_access until later. However, it may be turn out to be a bit messy, and it will increase code duplication.

[kgutwin]

As of at least version 1.15.0, this issue seems to be resolved...

$ mypy --version
mypy 1.15.0 (compiled: yes)
$ mypy x.py  # from the original issue report
x.py:6: note: Revealed type is "_io.TextIOWrapper[_io._WrappedBuffer]"
x.py:7: note: Revealed type is "_io.TextIOWrapper[_io._WrappedBuffer]"
x.py:12: note: Revealed type is "_io.TextIOWrapper[_io._WrappedBuffer]"
x.py:13: note: Revealed type is "_io.TextIOWrapper[_io._WrappedBuffer]"
Success: no issues found in 1 source file

[hauntsaninja]

Hm yeah, looks like it got fixed in a more recent typeshed sync: #17971

Jukka's self contained repro still reproduces, so we should keep this issue open

It could also maybe be worth adding a test case to typeshed