Fix validate_call binding issue with inheritance

[sydney-runkle]

Change Summary

Fix #8146

Checklist

• The pull request title is a good summary of the changes - it will be used in the changelog
• Unit tests for the changes exist
• Tests pass on CI
• Documentation reflects the changes where applicable
• My PR is ready to review, please add a comment including the phrase "please review" to assign reviewers

Selected Reviewer: @lig

[cloudflare-pages]

Deploying with    Cloudflare Pages

Latest commit:	6c7fd7f
Status:	✅  Deploy successful!
Preview URL:	https://6bcb3843.pydantic-docs2.pages.dev
Branch Preview URL:	https://fix-validate-call-issue.pydantic-docs2.pages.dev

View logs

[codspeed-hq]

CodSpeed Performance Report

Merging #8258 will not alter performance

_{Comparing fix-validate-call-issue (6c7fd7f) with main (667cd37)}

Summary

✅ 10 untouched benchmarks

[sydney-runkle]

@alexmojaki, what are your thoughts on this?

We've been debating a bit - on one hand the qualname approach is a bit dirty. On the other hand, we haven't yet figured out a clean approach (we tried a weakref cache, but couldn't get that working).

We want to have a solution that maintains the performance benefits of rebinding the wrapped function to an instance in general, while also catching this specific use case where we don't want to bind.

[sydney-runkle]

Please review

[alexmojaki]

ValidateCallWrapper says that it's a descriptor so that it "can be applied to instance methods, class methods, static methods, as well as normal functions." I don't understand this reasoning. In #8268 it looks like returning a plain function from @validate_call works fine*. The behaviour changes slightly, so maybe we could have a discussion about compatibility, but overall it feels like an improvement.

*(someone please rerun that failed pypy job though 😅)

[sydney-runkle]

Closing in favor of #8268