CohenArthur

I think this part should live in its own header + source files. We can then add attribute-specific functions such as the one to handle #[path] as written by @mathstuf...

The way the attribute works is that the attribute is gated: Without the `rustc_attrs` feature enabled, an error will be emitted by the compiler: `internal implementation detail`. I think that...

This is obviously until we can start compiling the standard library and want to look at `rustc_attrs` and all the builtin attributes it unlocks, created using the `rustc_attr!` macro in...

@liushuyu do you think we can close this one? Is there anything left you think you didn't handle in your PR?

Yes, but since there's a bunch of attributes hidden behind that feature I think it's fine to treat it separately. We also don't handle feature gatting at all at the...

> Isn't there an unstable feature that allows this? Oh fair point I didn't think of checking. I found [`arbitrary_enum_discriminant`](https://doc.rust-lang.org/beta/unstable-book/language-features/arbitrary-enum-discriminant.html) which implements it on top of adding discriminants to tuple...

I get this issue when trying to compile `gcc` (not `gccrs`) from source as well. Have you followed a specific tutorial or instructions when trying to cross-compile for SH4? I...

@Friedrich-S did you manage to fix that error? Sorry I forgot about it a bit. It's extremely weird, especially since this code is not specific to `gccrs` and is common...

By default, all identifiers referred to in a macro are expanded as-is, and are looked up at the macro's invocation site. This can lead to issues if a macro refers...

I'm all for factorisation - it's always easier when searching for where an error occurs to search in a central place, especially with format strings and so on. Plus, if...