fix: code hashes can now be be voted for in all code protocol states

[DSharifi]

closes #1589

[pbeza]

Left only a few optional nits. ✅

[kevindeforth]

Nice. So, right now, the big question is what to do with the TestSetup and TestSetupBuilder structs.

I like the proposed ergonomics and the user interface. .with_participant_count(participants) feels very nice.

I don't like that it relies so heavily on implemented functions and that it duplicates existing logic. For that, I think it would be good to leverage existing logic in our test_utils files.

The risk right now is that we steer towards maintaining two testing libraries for the same thing.

[kevindeforth]

I would suggest to take the interface you propose, as it's very nice, but implement it in the existing test_utils.rs files. This would achieve three things:

• it makes your nice interface accessible to other tests in the contract folder
• it unifies our test setup logic, so we only maintain one, not two
• it means when we generate a contract state, we generate it without relying on any voting paths, but can just insert the values in the struct.

[kevindeforth]

Thanks for the changes.

[netrome]

Looks great, thanks for adding the test case! 🙏