C API: EvalState -> nix_eval_state?

[roberth]

Is your feature request related to a problem? Please describe.

After using the bindings a bit, I've noticed that we currently have a mix of CPlusPlusLike identifiers and nix_c_api_identifiers for types.

Maybe the rule is that types with a direct C++ equivalent reuse that name? Even then it's probably not consistent now, and too arbitrary, as the C++ code can change over time.

My apologies if this has already been discussed.

Describe the solution you'd like

Switch to snake case for all types? E.g. nix_eval_state, nix_store_path, ...

We could add deprecated typedefs to help the existing bindings switch without breaking them.

Describe alternatives you've considered

• Leave as is -> confusing

• Change to CamelCase like the C++ code -> iirc we decided against this, so that it's obvious which API is used

Additional context

• (Towards) stable C bindings for libutil, libexpr #8699

Priorities

Add 👍 to issues you find important.

[jlesquembre]

I also think it makes sense to switch to snake_case.

I'm happy to open a PR for these changes if we agree on it. But I don't think we need to add deprecated typedefs if we do the change soon. Currently, we almost don't have any bindings, and we are introducing other breaking changes anyway.