Fix a false positive for class attribute typed with Final #10712
Conversation
Pierre-Sassoulas
added
False Positive 🦟
|
Do you have any thoughts on the expected messages Jacob ? |
|
|
||
| @dataclass | ||
| class Class: | ||
| class_snake_case_constant: Final[int] = 42 # [invalid-name] |
There was a problem hiding this comment.
This line should not raise [invalid-name], because Final attributes can still be re-assigned a new value that is different from the default value in __init__(). Final only protects the value after the instance is created.
Final is like const in C/C++. UPPER_CASE is like the marco which gets expanded during static interpretation of the code.
So UPPER_CASE is more immutable than Final.
So marking a dataclass attribute Final is not the sufficient condition for UPPER_CASE.
| @dataclass | ||
| class Class: | ||
| class_snake_case_constant: Final[int] = 42 # [invalid-name] | ||
| CLASS_UPPER_CASE_CONSTANT: Final[int] = 42 |
There was a problem hiding this comment.
This line should be OK. Stacking them together should not raise any [invalid-name].
| @dataclass | ||
| class Class: | ||
| class_snake_case_constant: Final[int] = 42 # [invalid-name] | ||
| CLASS_UPPER_CASE_CONSTANT: Final[int] = 42 |
There was a problem hiding this comment.
I suggest adding a just UPPER_CASE attribute test case like:
CLASS_UPPER_CASE_ATTRIBUTE: int = 42This line should not raise [invalid-name] here, because UPPER_CASE can be used alone to identify a static constant.
But later on if it gets re-assigned new value in the below method, then [invalid-name] should be raised there to ask for snake_case because it is no longer a constant.
| CLASS_UPPER_CASE_CONSTANT: Final[int] = 42 | ||
|
|
||
| def method(self) -> None: | ||
| method_snake_case_constant: Final[int] = 42 # [invalid-name] |
There was a problem hiding this comment.
I suggest adding a test case for
self.CLASS_UPPER_CASE_ATTRIBUTE = 36 # [invalid-name]
Because this attribute is named with UPPER_CASE letters, it should not be re-assigned a new value by any means. If it is re-assigned a new value, we can safely raise [invalid-name] on the spot and asking for snake_case names instead.
|
I'll aim to take a look this week at fixing this. |
Type of Changes
Description
Closes #10711