[ty] Infer TypeVarTuple arguments from variadic calls

[charliermarsh]

Summary

Follow-up to #27516.

Previously, we inferred Unknown for a TypeVarTuple appearing directly in a variadic parameter, even when its argument types and tuple shape were known:

def collect[*Ts](*args: *Ts) -> tuple[*Ts]: ...

def caller(i: int, s: str) -> None:
    reveal_type(collect())      # tuple[()]
    reveal_type(collect(i, s))  # tuple[int, str]

We now infer the complete argument sequence as a single tuple, preserving fixed prefixes and suffixes, starred tuple shapes, generic constructors, and both PEP 695 and legacy declarations.

Calls requiring joint inference through callbacks or other parameters, forwarded tuple unions, and conflicting contextual return types retain their existing behavior.

The number of passing typing-conformance files increases from 104 to 106.

[astral-sh-bot]

Typing conformance results improved 🎉

The percentage of diagnostics emitted that were expected errors increased from 96.97% to 97.35%. The percentage of expected errors that received a diagnostic held steady at 93.24%. The number of fully passing files improved from 104/133 to 106/133.

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	993	993	+0
False Positives	31	27	-4	⏬ (✅)
False Negatives	72	72	+0
Total Diagnostics	1066	1062	-4	⏬
Precision	96.97%	97.35%	+0.38%	⏫ (✅)
Recall	93.24%	93.24%	+0.00%
Passing Files	104/133	106/133	+2	⏫ (✅)

Test file breakdown

2 files altered

File	True Positives	False Positives	False Negatives	Status
generics_typevartuple_args.py	8	0 (-3) ✅	0	✅ Newly Passing 🎉
generics_typevartuple_callable.py	1	0 (-1) ✅	0	✅ Newly Passing 🎉
Total (all files)	993	27 (-4) ✅	72	106/133

False positives removed (4)

4 diagnostics

Test case	Diff
generics_typevartuple_args.py:29	-error[type-assertion-failure] Type `tuple[Unknown, ...]` does not match asserted type `tuple[int, str]`
generics_typevartuple_args.py:31	-error[type-assertion-failure] Type `tuple[Unknown, ...]` does not match asserted type `tuple[()]`
generics_typevartuple_args.py:32	-error[type-assertion-failure] Type `tuple[Unknown, ...]` does not match asserted type `tuple[int, str]`
generics_typevartuple_callable.py:50	-error[type-assertion-failure] Type `tuple[int | float | complex, *tuple[Unknown, ...]]` does not match asserted type `tuple[int | float | complex, str, int | float]`

[astral-sh-bot]

Memory usage report

Memory usage unchanged ✅

[astral-sh-bot]

ecosystem-analyzer results

No diagnostic changes detected ✅

Flaky changes detected. This PR summary excludes flaky changes; see the HTML report for details.

Full report with detailed diff (timing results)