Mypy pydantic plugin comparaison no error

[tyki6]

Initial Checks

• I confirm that I'm using Pydantic V2

Description

With mpy pydantic plugin enable when i enable strict equality pydantic comparaison not generate error when i write mistake error

Example Code

[tool.mypy]
plugins = "pydantic.mypy"
ignore_missing_imports = true
warn_return_any = true
warn_unreachable = true
warn_unused_configs = true
follow_imports = "normal"
show_column_numbers = true
strict_optional = true
warn_redundant_casts = true
pretty = false
strict = true
warn_unused_ignores = true
check_untyped_defs = true
disallow_untyped_calls = true
disallow_untyped_defs = true
disallow_untyped_decorators = false
strict_equality = true
[tool.pydantic-mypy]
init_forbid_extra = true
init_typed = true
warn_required_dynamic_aliases = true


```python
from pydantic import BaseModel
class User(BaseModel):
    username: str
user = User(username="test")
print(user == "test") # no error
print(user.username == [1,2,3]) # no error
print(user.username == "test") 

mypy .

I got:

Success: no issues found

Python, Pydantic & OS Version

pydantic version: 2.6.0
        pydantic-core version: 2.16.1
          pydantic-core build: profile=release pgo=trueC
               python version: 3.12.1 (main, Dec  7 2023, 20:45:44) [GCC 11.4.0]
                     platform: Linux-5.15.133.1-microsoft-standard-WSL2-x86_64-with-glibc2.35
             related packages: pydantic-settings-2.1.0 mypy-1.8.0 fastapi-0.109.0 typing_extensions-4.9.0 email-validator-2.1.0.post1
                       commit: unknown

[sydney-runkle]

@dmontagu,

Could you please take a look at this! Thanks :).

[dmontagu]

@tyki6 what version of mypy are you using?

I checked both mypy v1.8.0 and v1.1.1 in both, I get:

/Users/davidmontague/Library/Application Support/JetBrains/PyCharm2023.3/scratches/scratch.py:14:7: error: Non-overlapping equality check (left operand type: "str", right operand type: "list[int]")  [comparison-overlap]

on the line with user.username == [1,2,3], which you seemed to mark as # no error. That said, I also get an error on the previous line if I change it to a dataclass, but not with BaseModel.

I suspect the reason for this is that we have defined __eq__ method on BaseModel with other as Any. If we hide that __eq__ from the type-checker using if not typing.TYPE_CHECKING, the issue goes away, so that may be the best fix, I'll run that by the other maintainers and if they agree we can merge it. (Another alternative may be to change the type hint on the other argument to __eq__ but I'm more concerned that changing that may affect other users.