[ty] Support infer_variance for legacy TypeVar

[charliermarsh]

Summary

See: #24927 (comment).

[astral-sh-bot]

Typing conformance results improved 🎉

The percentage of diagnostics emitted that were expected errors increased from 88.79% to 89.34%. The percentage of expected errors that received a diagnostic increased from 84.63% to 84.82%. The number of fully passing files held steady at 84/134.

Summary

How are test cases classified?

Each test case represents one expected error annotation or a group of annotations sharing a tag. Counts are per test case, not per diagnostic — multiple diagnostics on the same line count as one. Required annotations (E) are true positives when ty flags the expected location and false negatives when it does not. Optional annotations (E?) are true positives when flagged but true negatives (not false negatives) when not. Tagged annotations (E[tag]) require ty to flag exactly one of the tagged lines; tagged multi-annotations (E[tag+]) allow any number up to the tag count. Flagging unexpected locations counts as a false positive.

Metric	Old	New	Diff	Outcome
True Positives	903	905	+2	⏫ (✅)
False Positives	114	108	-6	⏬ (✅)
False Negatives	164	162	-2	⏬ (✅)
Total Diagnostics	1067	1063	-4	⏬
Precision	88.79%	89.34%	+0.55%	⏫ (✅)
Recall	84.63%	84.82%	+0.19%	⏫ (✅)
Passing Files	84/134	84/134	+0

Test file breakdown

1 file altered

File	True Positives	False Positives	False Negatives	Status
generics_syntax_infer_variance.py	18 (+2) ✅	0 (-6) ✅	1 (-2) ✅	📈 Improving
Total (all files)	905 (+2) ✅	108 (-6) ✅	162 (-2) ✅	84/134

True positives added (2)

2 diagnostics

Test case	Diff
generics_syntax_infer_variance.py:15	+error[invalid-legacy-type-variable] A `TypeVar` cannot specify variance when `infer_variance=True`
generics_syntax_infer_variance.py:17	+error[invalid-legacy-type-variable] A `TypeVar` cannot specify variance when `infer_variance=True`

False positives removed (6)

6 diagnostics

Test case	Diff
generics_syntax_infer_variance.py:166	-error[invalid-assignment] Object of type `ShouldBeContravariant1[int | float]` is not assignable to `ShouldBeContravariant1[int]`
generics_syntax_infer_variance.py:28	-error[invalid-assignment] Object of type `ShouldBeCovariant1[int]` is not assignable to `ShouldBeCovariant1[int | float]`
generics_syntax_infer_variance.py:46	-error[invalid-assignment] Object of type `ShouldBeCovariant2[int]` is not assignable to `ShouldBeCovariant2[int | float]`
generics_syntax_infer_variance.py:55	-error[invalid-assignment] Object of type `ShouldBeCovariant3[int]` is not assignable to `ShouldBeCovariant3[int | float]`
generics_syntax_infer_variance.py:84	-error[invalid-assignment] Object of type `ShouldBeCovariant5[int]` is not assignable to `ShouldBeCovariant5[int | float]`
generics_syntax_infer_variance.py:95	-error[invalid-assignment] Object of type `ShouldBeCovariant6[int]` is not assignable to `ShouldBeCovariant6[int | float]`

[astral-sh-bot]

ecosystem-analyzer results

No diagnostic changes detected ✅

Full report with detailed diff (timing results)

[astral-sh-bot]

Memory usage report

Summary

Project	Old	New	Diff	Outcome
prefect	676.59MB	676.87MB	+0.04% (284.59kB)	⏫
flake8	47.78MB	47.78MB	-	✅
sphinx	260.02MB	260.02MB	-	✅
trio	117.07MB	117.07MB	-	✅

Significant changes

Click to expand detailed breakdown

prefect

Name	Old	New	Diff	Outcome
infer_expression_type_impl	7.41MB	7.46MB	+0.74% (55.76kB)	⏫
infer_expression_types_impl	55.89MB	55.94MB	+0.10% (55.63kB)	⏫
StaticClassLiteral<'db>::implicit_attribute_inner_	6.58MB	6.62MB	+0.57% (38.36kB)	⏫
infer_definition_types	85.73MB	85.77MB	+0.04% (32.53kB)	⏫
Type<'db>::member_lookup_with_policy_	15.77MB	15.80MB	+0.19% (31.12kB)	⏫
Type<'db>::class_member_with_policy_	16.85MB	16.88MB	+0.16% (27.77kB)	⏫
StaticClassLiteral<'db>::variance_of_	53.30kB	64.93kB	+21.82% (11.63kB)	⏫
all_negative_narrowing_constraints_for_expression	1.93MB	1.94MB	+0.42% (8.24kB)	⏫
all_narrowing_constraints_for_expression	5.99MB	6.00MB	+0.13% (8.24kB)	⏫
infer_scope_types_impl	55.39MB	55.40MB	+0.01% (5.74kB)	⏫
StaticClassLiteral<'db>::variance_of_::interned_arguments	47.88kB	50.84kB	+6.17% (2.95kB)	⏫
Type<'db>::apply_specialization_	3.61MB	3.62MB	+0.06% (2.18kB)	⏫
GenericAlias<'db>::variance_of_	40.72kB	42.29kB	+3.86% (1.57kB)	⏫
GenericAlias<'db>::variance_of_::interned_arguments	34.03kB	35.30kB	+3.72% (1.27kB)	⏫
when_constraint_set_assignable_to_owned_impl	1.53MB	1.53MB	-0.05% (736.00B)	⏫
...			16 more

[AlexWaygood]

This is obviously a huge improvement as-is, but do we know what the last remaining false negative is here?

[charliermarsh]

Oh I didn't notice it was just one more! I'll look.

[charliermarsh]

Oh no!

# vo4_4: ShouldBeCovariant4[int] = ShouldBeCovariant4[float](1.0)  # E

[charliermarsh]

(I'll put up a separate PR for that after confirming.)

[astral-sh-bot]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

Formatter (stable)

✅ ecosystem check detected no format changes.

Formatter (preview)

✅ ecosystem check detected no format changes.