RUF009 false positive for descriptor fields #4451
Comments
|
You can use flake8-bugbear.extend-immutable-calls to set your descriptor as immutable. The documentation could be explicit on the possibility to mark a call as immutable. |
|
That's almost certainly the "best we can do" for now, until we can reliably detect that a class is a descriptor across files. |
|
Thanks @JonathanPlasse. That looks like it will work for now. @charliermarsh, what about within a file? Many of my descriptor field classes are defined in the same file as the dataclass. |
|
Ruff did not have the type checking necessary to do this now. |
|
We might be able to do it based on references though, even without type-checking. In the linked example, if the file were |
|
@charliermarsh, should we add a tag for issues that could be improved with type checking ability? |
|
Yeah, we have an existing tag -- just added it. |
## Summary
Per the Python documentation, dataclasses are allowed to instantiate
descriptors, like so:
```python
class IntConversionDescriptor:
def __init__(self, *, default):
self._default = default
def __set_name__(self, owner, name):
self._name = "_" + name
def __get__(self, obj, type):
if obj is None:
return self._default
return getattr(obj, self._name, self._default)
def __set__(self, obj, value):
setattr(obj, self._name, int(value))
@DataClass
class InventoryItem:
quantity_on_hand: IntConversionDescriptor = IntConversionDescriptor(default=100)
```
Closes #4451.
Dataclasses permit descriptor fields (https://docs.python.org/3.10/library/dataclasses.html#descriptor-typed-fields) but RUF009 is complaining about function calls at class construction. This is similar to #4091, but I think sufficiently different to merit its own issue.
The text was updated successfully, but these errors were encountered: