Derived Version Structs

[npmccallum]

1. add a derive macro for deriving enumerations that are tagged as INTEGER.
2. use the new macro for pkcs10::Version.
3. add x509::Version.

[tarcieri]

This seems like you're somewhat needlessly defining another proc macro which is ultimately just a 1-argument wrapper for Enumerated.

Integer is really not very descriptive name and kind of ambiguous here: what it's doing is mapping integers to enums, just like Enumerated.

I'd suggest just making the Enumerated macro more flexible instead.

[npmccallum]

Add I said before, I already tried that. But Rust doesn't remove the type-level attribute, so it is passed through to the final AST and generates a compile error.

[tarcieri]

That must've been a bug in your code, as I've written many proc macros that do that without issue, including this one (see TypeAttrs)

[tarcieri]

This seems like an inextensible way to override the tag of an Enumerated, so why not just add a #[asn1(tag = "...")] argument instead that lets you override the ENUMERATED tag however you want? Then it could potentially work for more than just INTEGER.

[npmccallum]

@tarcieri I have implemented it with the attributes. However, the error I mentioned before has returned:

error: cannot find attribute `asn1` in this scope
 --> pkcs10/src/version.rs:9:3
  |
9 | #[asn1(type = "INTEGER")]
  |   ^^^^

This is because the derive macro outputs the asn1 attribute in the final AST. But Rust doesn't know about this attribute.

[tarcieri]

@npmccallum the problem is in the definition of the Enumerated proc macro:

https://github.com/RustCrypto/formats/blob/master/der/derive/src/lib.rs#L213-L218

#[proc_macro_derive(Enumerated)]
#[proc_macro_error]
pub fn derive_enumerated(input: TokenStream) -> TokenStream {
    let input = parse_macro_input!(input as DeriveInput);
    DeriveEnumerated::new(input).to_tokens().into()
}

You'll need to notate that it accepts an asn1 attribute. Change:

#[proc_macro_derive(Enumerated)]

to

#[proc_macro_derive(Enumerated, attributes(asn1))]

[npmccallum]

@tarcieri THANK YOU! You have no idea how much I was pulling my hair out on that one.

[npmccallum]

@tarcieri This is ready for final review now.

[tarcieri]

This is... less than ideal, but I have some ideas about how to refactor this along with attribute parsing in general which I can do in a followup PR.