[goodplay]

This is good news for us who want alternative implementations like gccrs. GCC (and software developers) can target static "epoch"s (c89, c99, c11, ...) rather than a moving target.

Hopefully, this will be enough to make development of GCC's rust front-end viable again.

[Manishearth]

FWIW the GCC frontend development seems to have started up again. I'm hopeful!

However, epoches don't really help alternate frontends. Rust 1.42 on epoch 2015 will still have all the features Rust 1.42 on the latest epoch has, except for the features that require an epoch shift (which will be a small handful)

But compiler versions already help here, targeting a compiler version gets you all of this.

I suspect if a compiler backend project nears completion there will be a lot more collaboration between rustc and the project in order to make evolution less painful.

[JoshTriplett]

> FWIW the GCC frontend development seems to have started up again. I'm hopeful!

I'm curious: any particular goals or results you hope to see out of the GCC frontend? Do they have any plans to share front-end code with rustc, such as the parser?

[Manishearth]

> any particular goals or results you hope to see out of the GCC frontend

- Support for weird architectures

- Potential for diverse double compiling

- Better crystallization of the language spec.

[steveklabnik]

One thing I'm hoping is that it brings support for some more obscure arches.

[JoshTriplett]

The ones that GCC supports and LLVM doesn't?

[steveklabnik]

Yup. There's a lot, in my understanding.