mrustc VS llvm-cbe

No, you don't. Existential proof: mrustc ignores lifetimes. Just flat out simply ignores. It changes some corner-cases related to HRBT, yet rustc compiled by mrustc works (that's BTW mrustc exist: to bootsrap the rustc compiler).

Incidentally C++ is the only way to bootstrap rust without rust today.

https://github.com/thepowersgang/mrustc

Well, there is mrustc[0], a Rust compiler that doesn't include a borrow-checker, so it's possible to compile (at least some versions of) Rust without a borrow checker, though it might not result in the most optimized code.

AFAIK there are some optimization like the infamous `noalias` optimization (which took several tries to get turned on[1]) that uses information established during borrow checking.

I'm also not sure what the relation with NLL (non-lexical lifetimes) is, where I would assume you would need at least a primitive borrow-checker to establish some information that the backend might be interested in. Then again, mrustc compiles Rust versions that have NLL features without a borrow-checker, so it's again probably more on the optimization side than being essential.

[0]: https://github.com/thepowersgang/mrustc

[1]: https://stackoverflow.com/a/57259339

> Maybe another memory safe language, but Rust has severe bootstrapping issues which is a hard sell for distros that care about source to binary transparency.

It is possible to bootstrap rustc from just GCC relatively easily, although it's a little bit time consuming.

You can use mrustc to bootstrap Rust 1.54: https://github.com/thepowersgang/mrustc

And from then you can go through each version all the way to the current 1.72. (Each new Rust version officially needs the previous one to compile.)

Have you tried this route : https://github.com/thepowersgang/mrustc ?

Mrustc supports Rust 1.54.0 today

There are three. The official one, mrustc (no borrow checker, but can essentially compile the official rustc) and GCC (can't really compile anything substantial yet). Only rustc is production-ready though.

That probably depends on what you mean by problematic. Having an ever increasing chain of dependencies isn’t the most desirable situation so there has been some work to trim the bootstrap chain. In 2018, when the blogpost I linked above was written, mrustc was used to bootstrap rust 1.19.0; now mrustc can bootstrap rust 1.54.0 so the chain to recent versions is much shorter than if all those intervening versions back through 1.19.0 needed to be built. https://github.com/thepowersgang/mrustc

So how does the LLVM C backend work then?

https://github.com/JuliaHubOSS/llvm-cbe

One alternative worth mentioning, though, would be the LLVM C Backend maintained by the Julia community.

If you drop "easily" and "human" (/s) from your requirements list, then the C backend for LLVM might work. Then you can choose any programming language you want that has LLVM 10-compatible frontend.

If you really must have something that compiles in C (e.g. for a platform where you only have a C compiler) there's an LLVM backend that outputs C code: https://github.com/JuliaComputingOSS/llvm-cbe

You can convert llvm bitcode to C and then use C compiler, there is such project https://github.com/JuliaComputingOSS/llvm-cbe .

I'm really surprised that the LLVM C backends have continually been resurrected then abandoned over the years. It's a good solution to this sort of thing and would enable a lot of cool stuff like Rust to weird embedded platforms. The most recent one is the Julia backend: https://github.com/JuliaComputingOSS/llvm-cbe

Check this project out: https://github.com/JuliaComputingOSS/llvm-cbe.

Well IMO it can definitely be rewritten in Julia, and to an easier degree than python since Julia allows hooking into the compiler pipeline at many areas of the stack. It's lispy an built from the ground up for codegen, with libraries like (https://github.com/JuliaSymbolics/Metatheory.jl) that provide high level pattern matching with e-graphs. The question is whether it's worth your time to learn Julia to do so.

You could also do it at the LLVM level: https://github.com/JuliaComputingOSS/llvm-cbe

For interesting takes on that, you can see https://github.com/JuliaLinearAlgebra/Octavian.jl which relies on loopvectorization.jl to do transforms on Julia AST beyond what LLVM does. Because of that, Octavian.jl beats openblas on many linalg benchmarks

You can try your luck with the "resurrected" C backend: https://github.com/JuliaComputingOSS/llvm-cbe

I don't understand why I see so many requests for LLVM-based languages to change around their backend or IR, that seems to be a huge amount of work for comparatively little benefit. The correct thing to do there is to just add support for those to LLVM.