Rust For The Linux Kernel Updated, Uutils As Rust Version Of Coreutils Updated Too

[moltonel]

I'm a rust fan and also see the value in autobuilds and tier. Also building rustcode with rustc has a much tighter implication than C++ with gcc (ownership check at compile time). Still I wouldn't see it as less trustworthy since gcc gets checked by hundreds of distros that build even more packages on a regular basis with it and therefore provide a huge quality control.

That's fair, Gcc gets a huge amount of testing. But after-the-fact testing is less satisfying than before-merge testing, and that's essentially what Rustc's tier list distinguishes. Rustc may see less code than Gcc but it's still an impressive amount, and LLVM is tested on much more than just rust code. Rustc and LLVM both have more code contributors per month than Gcc. And lastly there's the language difference, where it's much easier to be affected by a new "clever legal optimization" in C/C++ than in Rust.

All of this is subjective and multifaceted, there's no clear quality winner. But I worry less when installing the latest Rustc than the latest Gcc.

[moltonel]

You could take a look at rustc_codegen_gcc

I know and even sponsor that project. It should dampen the "can't use Rust for Linux because of $NICHE_PLATFORM" arguments, though I suspect the True Scotsmen purists will ask to wait for gccrs instead.

If rustup could distribute gcc + codegen backend for gcc, then the testing of Gcc in rustc could be much easier and might be one day tier 1, as every rust crate could use this new backend to run CI and CD.

It'll be a looong time before a codegen_gcc platform raises in the tiers. And even when it does, it won't stop the "can't use Rust on $PLATFORM (and therefore anywhere) because it's not Tier 1" arguments.

[NobodyXu]

I know and even sponsor that project. It should dampen the "can't use Rust for Linux because of $NICHE_PLATFORM" arguments, though I suspect the True Scotsmen purists will ask to wait for gccrs instead.

Sorry, I don't know what is "Scotsmen purists", but judging from your reply, it seems that they reject anything but GCC, which is fully GPL 2?

It'll be a looong time before a codegen_gcc platform raises in the tiers. And even when it does, it won't stop the "can't use Rust on $PLATFORM (and therefore anywhere) because it's not Tier 1" arguments.

Well, if that happens, that it's obviously that these kinds of arguments are either trolling, or are just people with their mind fixed to C with GCC and cannot accept anything else.

Fortunately, with many big corporates like Google adopting Rust and many being contributors to Linux, and the fact that Linus himself is open to adopting Rust in kernel, it is likely Rust can be used in the kernel.

[moltonel]

Sorry, I don't know what is "Scotsmen purists", but judging from your reply, it seems that they reject anything but GCC, which is fully GPL 2?

Well, if that happens, that it's obviously that these kinds of arguments are either trolling, or are just people with their mind fixed to C with GCC and cannot accept anything else.

See https://en.wikipedia.org/wiki/No_true_Scotsman For some people, true portabilty needs gccrs.

The fact is that "portability" has many definitions, levels and criterias. And to be fair, gccrs ticks some boxes that codegen_gcc doesn't (bootstrapping from a minimal C compiler, only one package to install, etc). Add to that a preference for GPL, the idea that Rust needs a spec and multiple implementations to be taken seriously, and you've got people earnestly wishing for gccrs. IMHO those problems are overstated, and are not worth the potential damage that an alternate compiler that can't keep with upstream can cause, but they're not necessarily trollish or unjustified.

[NobodyXu]

See https://en.wikipedia.org/wiki/No_true_Scotsman For some people, true portabilty needs gccrs.

The fact is that "portability" has many definitions, levels and criterias. And to be fair, gccrs ticks some boxes that codegen_gcc doesn't (bootstrapping from a minimal C compiler, only one package to install, etc). Add to that a preference for GPL, the idea that Rust needs a spec and multiple implementations to be taken seriously, and you've got people earnestly wishing for gccrs. IMHO those problems are overstated, and are not worth the potential damage that an alternate compiler that can't keep with upstream can cause, but they're not necessarily trollish or unjustified.

I also agree that having an alternative opensource compiler isn't good as it divides effort and create implementations that are slightly different, thus incompatible in a way.

[evasb]

I like Uutils, but re-licensing in a permissive license seems like a spit in the face of the original GNU authors.

[moltonel]

I like Uutils, but re-licensing in a permissive license seems like a spit in the face of the original GNU authors.

As much as GNU coreutil's licensing is a spit in the face of the original BSD authors. It's wrong to call this a re-licensing though: it's just an alternate implementation of a de-facto standard, like when Google (and GNU, with less success) reimplemented Sun's Java APIs.

[piotrj3]

I would say Rust on tier 2 platforms as long as you don't go into nightly is pretty safe. Nightly is pretty commonly used by Rust programmers so generally regressions/bugs are fixed before they go into release. Keep in mind tier 2 also runs commonly automated tests just it doesn't have so strict requirements as tier 1 targets do.