[red-knot] Update salsa

[MichaReiser]

Summary

Update Salsa to pull in salsa-rs/salsa#788 which fixes the, by now, famous access to field whilst the value is being initialized.

This PR also re-enables all tests that previously triggered the panic.

Test Plan

cargo test

[github-actions]

mypy_primer results

No ecosystem changes detected ✅

[codspeed-hq]

CodSpeed Performance Report

Merging #17320 will not alter performance

_{Comparing micha/update-salsa (09a8741) with main (144484d)}

Summary

✅ 32 untouched benchmarks

[MichaReiser]

Lol... Codspeed is confused. It thinks I acknowledged this regression a month ago.

[github-actions]

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.

[MichaReiser]

It looks like the main reason for the regression is that we now see more deep_verified_memo calls compared to before. I'm a bit confused about this because the old code should already always have reached the deep_verify_memo (unless it panicked). The only thing I can think of is that we now run deep_verify_memo for provisional memos where we always returned Changed before... but wouldn't that lead to a perf improvement instead of a regression?

[AlexWaygood]

Thanks!

[dcreager]

🎉 Thanks Micha!

[MichaReiser]

I think I may have identified the reason for the perf regression... salsa-rs/salsa#789

[MichaReiser]

TODO: Update the commit after my other PR is merged

[MichaReiser]

Yes, this was it. Now it's neutral.

[MichaReiser]

I'll merge this tomorrow. The salsa merge queue is too slow. If anyone wants this sooner. Feel free to update the Cargo.tomls with the commit hash from the salsa repo (tip of head)