refactor to remove add_to_module functions from generated code

[davidhewitt]

Related to #3989

@alex / @reaperhulk, want to try this branch with the cryptography CI and see if it fixes (or reduces) the problem?

This reshapes the generated macro output so that it only emits constants (no functions) and then we leave the implementations inside PyO3's impl_ module, as I suggested on that issue. This is probably better for binary size / compile times even if it doesn't help the coverage problem.

[codspeed-hq]

CodSpeed Performance Report

Merging #4009 will not alter performance

_{Comparing davidhewitt:add-to-module (9de7016) with main (dd17102)}

🎉 Hooray! codspeed-rust just leveled up to 2.4.0!

A heads-up, this is a breaking change and it might affect your current performance baseline a bit. But here's the exciting part - it's packed with new, cool features and promises improved result stability 🥳!
Curious about what's new? Visit our releases page to delve into all the awesome details about this new version.

Summary

✅ 68 untouched benchmarks

[alex]

This significantly improved the coverage situation. We still have a few things missing coverage:

• All invocations of pyo3::import_exception
• Exactly one pyclass
• One other bizzaro line (that's probably not related to pyo3)

[alex]

For clarity, I think it makes sense to merge this and then we can do follow up PRs resolving the other coverage issues.

[davidhewitt]

Great, will merge now!