This is early access, so there is great potential for finding internal compiler errors like we used to do back in the early 2000s for C++. How much fun that was. We kept on abusing templates and reported errors to debian upstream who then reported to gcc.
I’m quite unfamiliar with the implications. So GCC will support compiling rust which is cool. You’re saying that might find new bugs in GCC which is good for GCC.
Do you think there will be advantages to current rust users? Having more supported compilers is a good thing. Having competing compiler implementation might help for perf somehow? Anything else?
It also means support for CPU targets that LLVM doesn’t support.
While that may sound a bit niche, it makes Rust relevant for a lot of embedded devices, developers (including me) and cross-platform software (such as Linux). Like adding a missing piece of the puzzle, I’m sure it will have synergistic effects on everything around it.
For merely getting obscure CPU support there is another project that uses the current rustc implementation and adds GCC only as a back-end: https://github.com/rust-lang/rustc_codegen_gcc
The gccrs project is a complete rewrite of Rust from scratch.
We are probably going to get outdated LTS gccrs binaries that have security patches applied to them. This is the way many large companies work. So I suspect this will make Rust more accessible for them.
This is early access, so there is great potential for finding internal compiler errors like we used to do back in the early 2000s for C++. How much fun that was. We kept on abusing templates and reported errors to debian upstream who then reported to gcc.
I’m quite unfamiliar with the implications. So GCC will support compiling rust which is cool. You’re saying that might find new bugs in GCC which is good for GCC.
Do you think there will be advantages to current rust users? Having more supported compilers is a good thing. Having competing compiler implementation might help for perf somehow? Anything else?
It also means support for CPU targets that LLVM doesn’t support.
While that may sound a bit niche, it makes Rust relevant for a lot of embedded devices, developers (including me) and cross-platform software (such as Linux). Like adding a missing piece of the puzzle, I’m sure it will have synergistic effects on everything around it.
For merely getting obscure CPU support there is another project that uses the current
rustcimplementation and adds GCC only as a back-end: https://github.com/rust-lang/rustc_codegen_gccThe gccrs project is a complete rewrite of Rust from scratch.
Thanks that makes sense.
We are probably going to get outdated LTS gccrs binaries that have security patches applied to them. This is the way many large companies work. So I suspect this will make Rust more accessible for them.
Did they add a normal/fat stdlib?