Dataclasses erroring when descriptor has mutable default

[thejcannon]

Bug report

Bug description:

Running the following code:

from dataclasses import dataclass

class Descriptor:
    def __set_name__(self, owner, name: str) -> None:
        self.name = name

    def __get__(self, instance, owner) -> list[str]:
        if instance is None:
            return []
        return instance.__dict__[self.name]

    def __set__(self, instance, value: list[str]) -> None:
        instance.__dict__[self.name] = value


@dataclass()
class TestFoo:
    apple: Descriptor = Descriptor()

print(TestFoo().apple)

results in:

  File "/opt/homebrew/Cellar/python@3.13/3.13.3/Frameworks/Python.framework/Versions/3.13/lib/python3.13/dataclasses.py", line 860, in _get_field
    raise ValueError(f'mutable default {type(f.default)} for field '
                     f'{f.name} is not allowed: use default_factory')
ValueError: mutable default <class 'list'> for field apple is not allowed: use default_factory

CPython versions tested on:

3.13

Operating systems tested on:

macOS

[thejcannon]

I'm not sure what to think about this one, since this code is (to dataclass) indistinguishable from:

class Descriptor:
    default = []

    def __set_name__(self, owner, name: str) -> None:
        self.name = name

    def __get__(self, instance, owner) -> list[str]:
        if instance is None:
            return self.default
        return instance.__dict__[self.name]

    def __set__(self, instance, value: list[str]) -> None:
        instance.__dict__[self.name] = value

which would be a bug.

---

At the very least the error message isn't helpful 😓

[sobolevn]

Will this work for you?

from dataclasses import dataclass, field

class Descriptor:
    def __set_name__(self, owner, name: str) -> None:
        self.name = name

    def __get__(self, instance, owner) -> list[str]:
        if instance is None:
            return field(default_factory=list)  # <- returning a `field` object
        return instance.__dict__[self.name]

    def __set__(self, instance, value: list[str]) -> None:
        instance.__dict__[self.name] = value


@dataclass()
class TestFoo:
    apple: Descriptor = Descriptor()

print(TestFoo().apple)

[ericvsmith]

I’m not sure how the message could be improved, but I’m open to suggestions. Detecting this case is the issue.

[thejcannon]

Oh specifically the message suggesting using default_factory when I'm not using a field 🙃

[thejcannon]

Returning a field is wonderfully clever. I'll have to test that 😁