use vectorcall calling convention for performance

[davidhewitt]

Change Summary

Uses Python's "vectorcall" calling convention instead of PyAny::call1 to avoid creating intermediate tuples. Only works on supported Python versions. (3.9+, I think, let's confirm with CI.)

In the long term I'd like to add a feature to upstream PyO3 which does this and also supports keyword arguments. Seems like we don't generally use keyword arguments so not relevant to us here. Probably also with a more generic interface than just a slice-of-any as input.

Related issue number

N/A

Checklist

• Unit tests for the changes exist
• Documentation reflects the changes where applicable
• Pydantic tests pass with this pydantic-core (except for expected changes)
• My PR is ready to review, please add a comment including the phrase "please review" to assign reviewers

[codecov]

Codecov Report

Merging #792 (ba4fbe7) into main (9b540d0) will decrease coverage by 0.13%.
The diff coverage is 78.64%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #792      +/-   ##
==========================================
- Coverage   93.53%   93.40%   -0.13%     
==========================================
  Files         100      101       +1     
  Lines       14376    14441      +65     
  Branches       25       25              
==========================================
+ Hits        13446    13489      +43     
- Misses        924      946      +22     
  Partials        6        6              

Impacted Files	Coverage Δ
src/lib.rs	100.00% <ø> (ø)
src/py_vectorcall.rs	52.17% <52.17%> (ø)
src/input/input_python.rs	98.11% <100.00%> (+0.01%)	⬆️
src/serializers/filter.rs	96.10% <100.00%> (+0.03%)	⬆️
src/serializers/infer.rs	96.94% <100.00%> (ø)
src/serializers/type_serializers/format.rs	91.11% <100.00%> (ø)
src/serializers/type_serializers/function.rs	96.22% <100.00%> (+0.08%)	⬆️
src/url.rs	98.30% <100.00%> (ø)
src/validators/function.rs	90.83% <100.00%> (+0.17%)	⬆️
src/validators/union.rs	81.57% <100.00%> (ø)
... and 1 more

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f38bea6...ba4fbe7. Read the comment docs.

[codspeed-hq]

CodSpeed Performance Report

Merging #792 will degrade performances by 10.5%

_{Comparing dh/vectorcall (ba4fbe7) with main (f38bea6)}

Summary

❌ 1 regressions
✅ 125 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	dh/vectorcall	Change
❌	test_dont_raise_error_no_info	22.6 µs	25.3 µs	-10.5%

[davidhewitt]

This doesn't seem to meaningfully improve performance, but I'll leave it here for the moment as I found it was helpful in decimal validation in #763.

[davidhewitt]

Correction: it makes barely any difference even in #763 so I'll close this for now.