Re: Librsvg 2.40.20 is released and is the last release in the 2.40.x series

On Sat, Dec 16, 2017 at 8:17 PM, Michael Biebl <mbiebl gmail com> wrote:
Our porters, from what I know, tried to bootstrap rust on various
architectures but run into multiple issues.
A few of them you can see at

If you need more details, I can ask them. All I know that this
situation is basically unchanged for a long time already.

I think this is something only arch teams can work on. I don't have
access to mips or s390 machines.

Having worked with arch teams in Gentoo, I understand that they're all
overworked and the work is slow because the machines have almost no
resources, but the work has to be done.

Probably doesn't help that rustc considers anything non i386/amd64 as
non-first tier supported.

This is a bit of a misunderstanding. What "tier-1 support" means is
that their full CI is run for all supported architectures and OSes for
every PR, and then that PR is merged on top of the tip and the CI is
is run on that again. It's a very comprehensive level of of testing,
and aiui it also includes performance regression testing and so on.
It's basically impossible to run CI on that scale for an organisation
like Rust for fringe arches like s390/mips/m68k/etc, so they cannot
advertise it.

However, I have talked to upstream Rust folks in person and they have
always said that they welcome patches to improve arch support, and
they would really like it if distros ran CI for them as often as they
can and report regressions early in the release cycle.

Effectively, we'd have the same situation with Rust that we have for
all other packages in existence, but with a more welcoming community
behind it. The problem is that of bootstrapping the process, and that
can only happen once packages actually requiring Rust.

I think Debian should start using librsvg 2.41 and arch teams should
just deal with that reality.


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]