Fail to handle property and init-only field with same name

[Henkhogan]

Environment data

• Language Server version: v2022.12.20
• OS and version: Ubuntu 22.10
• Python version (& distribution if applicable, e.g. Anaconda): 3.11.0

Code Snippet

from dataclasses import dataclass, field, InitVar

@dataclass(slots=True)
class MyClass:
    foo: InitVar[str]  # type: ignore
    _foo: str = field(init = False)

    def __post_init__(self, foo):
        self.foo = foo

    @property
    def foo(self) -> str:
        return self._foo

    @foo.setter
    def foo(self, value: str):
        if not isinstance(value, str):
            raise TypeError(f'foo attribute must be of type: {str} but received: {value}')
        self._foo = value


c = MyClass(foo = 'foostring')





_ = c.foo

[erictraut]

This behavior is correct. You are attempting to redefine the symbol foo in an incompatible manner. Not surprisingly, mypy generates the same error here.

You should use different names for the init variable and the property if you want this to work with a static type checker.

[Henkhogan]

Well I think it should not be too difficult to handle that since it's already detected that we have an init-only field, which could be completely ignored for each access against instance attributes.

However if that's an issue that should rather be fixed in mypy than in pylance, I'll address it there

[erictraut]

This violates the Python type system, so there's nothing to "fix" in mypy. A symbol is not allowed to have two conflicting declarations. In your example, you've declared foo as being both an init variable and a property. The job of a type checker is to detect type violations — including conflicting type declarations. Both pylance and mypy are correct in flagging this as a violation. You'll need to choose non-conflicting names for the init variable and the property if you want this to work with a static type checker.

[erictraut]

If you're not interested in static type checking, you can set python.analysis.typeCheckingMode to "off", and these errors will not be reported by pylance.

[Henkhogan]

This violates the Python type system, so there's nothing to "fix" in mypy. A symbol is not allowed to have two conflicting declarations. In your example, you've declared foo as being both an init variable and a property. The job of a type checker is to detect type violations — including conflicting type declarations. Both pylance and mypy are correct in flagging this as a violation. You'll need to choose non-conflicting names for the init variable and the property if you want this to work with a static type checker.

@erictraut I am aware, but I don't see a real conflict here. What I see is a pseudo-conflict that will not be present at runtime. In my opinion the special characteristic of the init-only field is not sufficiently considered. Quoting the dataclass docs

If a field is an InitVar, it is considered a pseudo-field called an init-only field. As it is not a true field, it is not returned by the module-level fields() function. Init-only fields are added as parameters to the generated init() method, and are passed to the optional post_init() method. They are not otherwise used by dataclasses.

[Henkhogan]

Duplicates #python/mypy#12046