[ty] Emit invalid type form for PEP 695 aliases by charliermarsh · Pull Request #24224 · astral-sh/ruff

charliermarsh · 2026-03-26T23:30:59Z

Summary

astral-sh-bot · 2026-03-26T23:32:55Z

Typing conformance results

No changes detected ✅

Current numbers

The percentage of diagnostics emitted that were expected errors held steady at 86.59%. The percentage of expected errors that received a diagnostic held steady at 80.96%. The number of fully passing files held steady at 68/132.

astral-sh-bot · 2026-03-26T23:33:47Z

Memory usage report

Memory usage unchanged ✅

astral-sh-bot · 2026-03-26T23:36:38Z

`ecosystem-analyzer` results

Lint rule	Added	Removed	Changed
`invalid-await`	40	0	0
`invalid-return-type`	1	0	0
Total	41	0	0

Changes in flaky projects detected. Raw diff output excludes flaky projects; see the HTML report for details.

Full report with detailed diff (timing results)

charliermarsh · 2026-03-27T00:04:54Z

crates/ty_python_semantic/resources/mdtest/pep695_type_aliases.md

+type Bad6 = NotRequired[int]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad7 = ReadOnly[int]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad8 = InitVar[int]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+# error: [invalid-type-form] "`InitVar` may not be used without a type argument"


This isn't great, we don't show this for bad10: TypeAlias = InitVar.

Fixed by sharing the codepaths.

AlexWaygood · 2026-03-27T12:46:32Z

I think there's a much simpler solution here, give me a second.

AlexWaygood · 2026-03-27T13:29:28Z

See #24242. Ideally I think we would do our PEP-613 validation more similarly to how we do our PEP-695 validation, but that throws up other issues...

charliermarsh · 2026-03-27T14:04:00Z

Just to clarify, you prefer #24242?

carljm

Functionality looks good here, just some nits about how we name and comment things.

carljm · 2026-03-27T03:31:45Z

crates/ty_python_semantic/resources/mdtest/pep695_type_aliases.md

+type Bad1 = ClassVar[str]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad2 = ClassVar  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad3 = Final[int]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad4 = Final  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad5 = Required[int]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad6 = NotRequired[int]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad7 = ReadOnly[int]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad8 = InitVar[int]  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"
+type Bad9 = InitVar  # error: [invalid-type-form] "Type qualifiers are not allowed in type alias definitions"


Do we already have equivalent test for TypeAlias aliases?

carljm · 2026-03-27T14:00:26Z

crates/ty_python_semantic/src/types/infer/builder/annotation_expression.rs

+/// Controls the top-level semantics used when interpreting an annotation expression.
+///
+/// PEP 695 type-alias values mostly reuse annotation-expression inference, but a few
+/// top-level special cases such as bare `InitVar` should behave like alias RHSs rather


I don't know what "behave like alias RHSs" means here. Don't we just mean to say that most type qualifier special forms are prohibited in PEP 695 alias right hand sides?

carljm · 2026-03-27T14:04:43Z

crates/ty_python_semantic/src/types/infer/builder/annotation_expression.rs

+    /// Annotation semantics for `x: TypeAlias = ...`.
+    Pep613TypeAlias,


This is misleading if treated as parallel to TypeAliasValue. This is not used for "right hand side of a PEP 613 type alias", it's used for annotated assignments, where an annotation of TypeAlias is allowed (vs other annotation contexts where it is not).

Not clear to me entirely if we should name this AnnotatedAssignment or PEP613AnnotationAllowed -- the former emphasizes when it is used, the latter emphasizes what difference it makes. I guess I kind of learn toward the former?

Suggested change

/// Annotation semantics for `x: TypeAlias = ...`.

Pep613TypeAlias,

/// The annotation of an annotated assignment statement.

AnnotatedAssignment,

carljm · 2026-03-27T14:07:00Z

crates/ty_python_semantic/src/types/infer/builder/annotation_expression.rs

+    /// Annotation semantics for `type X = ...`.
+    TypeAliasValue,


Two nits on naming here: "TypeAlias" suggests PEP 613, if we don't say PEP695, and "Value" mildly suggests "as a value expression", which is not what we're doing here.

Suggested change

/// Annotation semantics for `type X = ...`.

TypeAliasValue,

/// Annotation semantics for the right-hand-side of `type X = ...`.

PEP695AliasRHS,

carljm · 2026-03-27T14:07:48Z

crates/ty_python_semantic/src/types/infer/builder/annotation_expression.rs

+    }
+
+    /// Infer the type of a PEP 695 type alias value.
+    pub(super) fn infer_type_alias_value_expression(


Suggested change

pub(super) fn infer_type_alias_value_expression(

pub(super) fn infer_pep_695_alias_rhs(

carljm · 2026-03-27T14:12:16Z

Oops, had this open in a tab on "changes" view from last night and missed Alex's comments.

charliermarsh · 2026-03-27T14:22:51Z

Yeah Alex and I have dueling solutions now. Unfortunately I have to assume his is better lol.

astral-sh-bot bot added the ty Multi-file analysis & type inference label Mar 26, 2026

charliermarsh commented Mar 27, 2026

View reviewed changes

charliermarsh marked this pull request as ready for review March 27, 2026 00:54

charliermarsh requested review from AlexWaygood, carljm, dcreager, ibraheemdev and sharkdp as code owners March 27, 2026 00:54

astral-sh-bot bot assigned carljm Mar 27, 2026

charliermarsh added 2 commits March 26, 2026 20:58

[ty] Emit invalid type form for PEP 695 aliases

5a0b867

Share code

f1293e6

charliermarsh force-pushed the charlie/tf branch from c6cc70b to f1293e6 Compare March 27, 2026 00:59

AlexWaygood mentioned this pull request Mar 27, 2026

[ty] Ban type qualifiers in PEP-695 type aliases #24242

Merged

carljm approved these changes Mar 27, 2026

View reviewed changes

charliermarsh closed this Mar 27, 2026

		/// Annotation semantics for `x: TypeAlias = ...`.
		Pep613TypeAlias,

	pub(super) fn infer_type_alias_value_expression(
	pub(super) fn infer_pep_695_alias_rhs(

Conversation

charliermarsh commented Mar 26, 2026

Summary

Uh oh!

astral-sh-bot bot commented Mar 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

No changes detected ✅

Uh oh!

astral-sh-bot bot commented Mar 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Memory usage report

Uh oh!

astral-sh-bot bot commented Mar 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

ecosystem-analyzer results

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlexWaygood commented Mar 27, 2026

Uh oh!

AlexWaygood commented Mar 27, 2026

Uh oh!

charliermarsh commented Mar 27, 2026

Uh oh!

carljm left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

carljm commented Mar 27, 2026

Uh oh!

charliermarsh commented Mar 27, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

astral-sh-bot bot commented Mar 26, 2026 •

edited

Loading

astral-sh-bot bot commented Mar 26, 2026 •

edited

Loading

astral-sh-bot bot commented Mar 26, 2026 •

edited

Loading

`ecosystem-analyzer` results