[ty] relax the attribute narrowing condition#22440
Merged
carljm merged 6 commits intoastral-sh:mainfrom Feb 6, 2026
Merged
Conversation
Typing conformance resultsNo changes detected ✅ |
|
|
| Lint rule | Added | Removed | Changed |
|---|---|---|---|
unresolved-attribute |
187 | 20 | 0 |
unsupported-operator |
3 | 86 | 50 |
not-subscriptable |
1 | 100 | 0 |
possibly-missing-attribute |
26 | 33 | 35 |
invalid-assignment |
22 | 4 | 5 |
invalid-argument-type |
16 | 5 | 9 |
no-matching-overload |
21 | 2 | 0 |
unused-type-ignore-comment |
3 | 6 | 0 |
invalid-parameter-default |
0 | 0 | 7 |
invalid-return-type |
1 | 1 | 2 |
not-iterable |
0 | 0 | 2 |
possibly-unresolved-reference |
0 | 2 | 0 |
call-non-callable |
1 | 0 | 0 |
missing-argument |
1 | 0 | 0 |
| Total | 282 | 259 | 110 |
mtshiba
force-pushed
the
relax-attribute-narrowing
branch
from
b6d09c3 to
8238ad4
Compare
Collaborator
Author
|
The mypy_primer results look reasonable. |
mtshiba
requested review from
AlexWaygood,
carljm,
dcreager and
sharkdp
as code owners
carljm
approved these changes
Feb 4, 2026
Contributor
carljm
left a comment
carljm
left a comment
There was a problem hiding this comment.
Thank you! The logic here looks solid. Some suggested renamings (or non-renamings) and comment updates. One question I don't quite understand about why the existing exception for Unknown wasn't effective.
crates/ty_python_semantic/resources/mdtest/narrow/conditionals/nested.md
Outdated
Show resolved
Hide resolved
crates/ty_python_semantic/resources/mdtest/narrow/assignment.md
Outdated
Show resolved
Hide resolved
Contributor
|
Have you done any analysis of the ecosystem results here? |
Co-authored-by: Carl Meyer <[email protected]>
Co-authored-by: Carl Meyer <[email protected]>
Co-Authored-By: Carl Meyer <[email protected]>
Collaborator
Author
The latest results show that this PR has removed many more false positive errors. Many of the newly added errors are caused by dynamically adding attributes to objects. |
Contributor
|
Looks great! BTW if your PR is reviewed with a green checkmark (approved), you should feel free to merge it yourself after addressing any comments, unless you feel there is something in the comments that needs more discussion, or the changes are large and you want re-review (then you can request re-review). |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixes astral-sh/ty#2134
This PR is based on the "pragmatism" that narrowing should only be disabled when an attribute is a data descriptor, and should be left as is when it might be a data descriptor.
I believe this is somewhat justified given the fact that the current behavior is not strictly safe either.
This code doesn't throw an error, and there's nothing inherently unsound about it, but the narrowing result is still incorrect.
Even if the class attribute is
Undefined, there is the possibility of dynamic attribute additions or additions in subclasses (and this issue persists even if we don't mark implicit attribute types withUnknown |).But if we take this into account, this kind of narrowing of implicit attributes becomes completely inapplicable!
I'd like to see the results of mypy_primer and decide whether this idea makes sense.
Test Plan
mdtest updated