Followup changes to fix ruff & pyright warnings #203
Conversation
… about mutable classvar, subclass Statement in test_flake8_trio so the original Statement is properly hashable.
tests/test_flake8_trio.py
Outdated
| ) | ||
|
|
||
| def __hash__(self) -> int: | ||
| raise NotImplementedError |
There was a problem hiding this comment.
That could cause objects that are equal to have different hashes (because -1 is a wildcard value), which is Not Allowed ™️
https://docs.python.org/3/reference/datamodel.html#object.__hash__
Rereading it now, it turns out I misread that originally and I didn't have to complicate things as much as I did. I originally read it as __eq__ objects had to have the same hash, and objects with same hash value had to be equal to eachother...
So I could implement __hash__() with return hash(tuple(self.name, self.lineno))
tests/test_flake8_trio.py
Outdated
| @@ -50,6 +50,22 @@ | |||
| class ParseError(Exception): ... | |||
|
|
|||
|
|
|||
| class ColumnAgnosticStatement(Statement): | |||
| def __eq__(self, other: object) -> bool: | |||
There was a problem hiding this comment.
Probably worth a comment explaining what this class is, and why we override __eq__ and __hash__ (and what the semantics of this class are).
There was a problem hiding this comment.
ended up reverting stuff and implementing a simple __hash__ :)
Statementto add__eq__, so it's properly hashable elsewhere.error_codesis now annotated asMapping, silencing RUF012Barely needs much of a review. @thejcannon @aneeshusa