[ty] Bind typing.Self in class attributes and assignment#23108
Merged
[ty] Bind typing.Self in class attributes and assignment#23108
typing.Self in class attributes and assignment#23108Conversation
Typing conformance results improved 🎉The percentage of diagnostics emitted that were expected errors increased from 82.63% to 83.14%. The percentage of expected errors that received a diagnostic held steady at 74.24%. Summary
False positives removedDetails
|
|
Merging this PR will not alter performance
Comparing 
Footnotes
|
|
| Lint rule | Added | Removed | Changed |
|---|---|---|---|
unresolved-attribute |
934 | 16 | 7 |
invalid-argument-type |
15 | 70 | 19 |
possibly-missing-attribute |
58 | 13 | 0 |
invalid-return-type |
2 | 41 | 3 |
invalid-assignment |
2 | 22 | 10 |
invalid-type-form |
33 | 0 | 0 |
invalid-type-arguments |
0 | 18 | 0 |
unused-type-ignore-comment |
0 | 3 | 0 |
redundant-cast |
2 | 0 | 0 |
| Total | 1,046 | 183 | 39 |
charliermarsh
force-pushed
the
charlie/self-test
branch
from
1019f08 to
c086f6f
Compare
charliermarsh
changed the title
typing.Self in class attributes and assignmenttyping.Self in class attributes and assignment
charliermarsh
requested review from
AlexWaygood,
carljm,
dcreager and
sharkdp
as code owners
Contributor
|
Just adding a note that the 992 new false positives on zulip are due to missing django support. |
sharkdp
requested changes
Feb 6, 2026
Contributor
sharkdp
left a comment
sharkdp
left a comment
There was a problem hiding this comment.
Finishing my first review here, as I would first like to understand if we really need the supports_self_binding filtering.
Member
Author
|
For completeness: I also had #22487 which was written by Claude and iterated-on over the course of a few days, but I found it hard to understand and this is my attempt to simplify the solution. |
charliermarsh
force-pushed
the
charlie/self-test
branch
from
e3d8129 to
2c9e780
Compare
sharkdp
requested changes
Feb 10, 2026
Contributor
sharkdp
left a comment
sharkdp
left a comment
There was a problem hiding this comment.
I have some more questions, but the general direction definitely looks good — thank you for the updates.
| self_typevar_owner_class_literal(db, other), | ||
| ) { | ||
| return left == right && self.paramspec_attr(db) == other.paramspec_attr(db); | ||
| } |
Contributor
There was a problem hiding this comment.
With this change, we will now also consider Self@method1 and Self@method2 the same typevar. Can this cause any problems?
Member
Author
There was a problem hiding this comment.
I think this is correct, e.g. given:
class Foo:
def method1(self) -> Self: ...
def method2(self) -> Self: ...
def combined(self) -> Self:
if condition:
return self.method1() # Self@method1
else:
return self.method2() # Self@method2You want Self@method1 | Self@method2 to simplify to Self.
Contributor
There was a problem hiding this comment.
I don't think we generate the type Self@method1 | Self@method2 here. Inferring the type for self.method1() will already bind Self@method1 and return Foo. This code works fine on main.
Interested in @dcreager's opinion here.
Member
Author
There was a problem hiding this comment.
(Increasingly think you're right, and that this was introduced in a prior iteration but subsequently made redundant. Going to try removing.)
Contributor
There was a problem hiding this comment.
Looks like it works, cool! In that case — no dcreager opinion needed 😄. Let's merge!
charliermarsh
force-pushed
the
charlie/self-test
branch
from
75c5ea3 to
58faf6f
Compare
Memory usage reportSummary
Significant changesClick to expand detailed breakdownprefect
sphinx
trio
flake8
|
charliermarsh
force-pushed
the
charlie/self-test
branch
from
f42072c to
4e48939
Compare
charliermarsh
force-pushed
the
charlie/self-test
branch
from
4e48939 to
4d8f4e0
Compare
sharkdp
approved these changes
Feb 11, 2026
Comment on lines
+665
to
+669
| class Linkable(Protocol): | ||
| next_node: Self | ||
|
|
||
| def advance(self) -> Self: | ||
| return self.next_node |
Contributor
There was a problem hiding this comment.
This passes on main as well. But if you try to access the attribute like this:
def _(linkable: Linkable):
reveal_type(linkable.next_node)we get @Todo(type[T] for protocols). We don't have to fix this TODO here, but I'm a bit surprised that other attributes can be accessed without problems (see below), so this might be worth investigating?
class Linkable(Protocol):
next_node: Self
other_attribute: int
def _(linkable: Linkable):
reveal_type(linkable.next_node) # revealed: @Todo(type[T] for protocols)
reveal_type(linkable.other_attribute) # revealed: intI think we should add a test like
Suggested change
| class Linkable(Protocol): | |
| next_node: Self | |
| def advance(self) -> Self: | |
| return self.next_node | |
| class Linkable(Protocol): | |
| next_node: Self | |
| def advance(self) -> Self: | |
| return self.next_node | |
| def _(l: Linkable) -> None: | |
| # TODO: Should be `Linkable` | |
| reveal_type(l.next_node) # revealed: @Todo(type[T] for protocols) |
| self_typevar_owner_class_literal(db, other), | ||
| ) { | ||
| return left == right && self.paramspec_attr(db) == other.paramspec_attr(db); | ||
| } |
Contributor
There was a problem hiding this comment.
I don't think we generate the type Self@method1 | Self@method2 here. Inferring the type for self.method1() will already bind Self@method1 and return Foo. This code works fine on main.
Interested in @dcreager's opinion here.
charliermarsh
force-pushed
the
charlie/self-test
branch
from
4d8f4e0 to
d883f6b
Compare
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
Closes astral-sh/ty#1124.