Re: rust 1.51.0 for 32-bit earmv7hf/9.0

[nia <nia%NetBSD.org@localhost>]

On Tue, May 18, 2021 at 05:13:30PM +0200, Havard Eidnes wrote:
> If this (the rust-bin modification) is deemed useful, I'll go
> ahead and commit what I have (and, yes, I'll let pkglint chew on
> it before I do :)

Oops, I already committed it. Hope I haven't stepped on anyone's
toes here.

> There appears to be work afoot to get "mrustc" (apparently
> written in C++) suitable for bootstrapping a full rust build
> without having to resort to a binary bootstrap, but we're not
> quite there yet.

I've been interested in mrustc, it's not a cure-all to all
of our rust-related problems (it appears to only really be suitable
for building rust itself), but should at least make the chain of trust
more verifiable for people who care about such things.
It does build on NetBSD without patching already!

It might be possible for us to contort mrustc to build other things,
but we'll probably need a separate package for cargo (mrustc
includes its own "minicargo", but claims it only builds enough
to build rustc itself), and stdlib. I wouldn't advise this
except on architectures where "mainline" rustc has absolutely
no chance, though of course it'd likely be another option
for RUST_TYPE.

Notably, mrustc doesn't have good error messages and
simply works by compiling rust to C.