Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.debian.devel > #112155

Re: About i386 support

From Maite Gamper <victor@wenzeslaus.de>
Newsgroups linux.debian.devel
Subject Re: About i386 support
Date 2024-06-14 13:20 +0200
Message-ID <IPd8l-2Sf3-3@gated-at.bofh.it> (permalink)
References <IPcP0-2RTn-7@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


Hello,

On 14.06.24 12:53, rhys@neoquasar.org wrote:
> Then it's not a problem in the first place. If you can't reproduce a 
> bug with a reasonable effort, then it is unconfirmed and you can stop 
> worrying about it. A bug that can't be reproduced, effectively doesn't 
> exist.
>
> That's not a reason to stop supporting an entire architecture. That's 
> a troubleshooting decision that you would make on any architecture.
>
> Sent from my mobile device.
>
> ------------------------------------------------------------------------
> *From:* Luca Boccassi <bluca@debian.org>
> *Sent:* Friday, June 14, 2024 04:39
> *To:* debian-devel@lists.debian.org
> *Subject:* Re: Re: About i386 support
>
> > I've tried reproducing the daemon-reload bug report, unless I missed
> > something
> > obvious, daemon-reload works on my T2300, the TM Efficeon, and the
> > pre-SSE2
> > Pentium 3 (mobile) that I have. I could try running it on an original
> > Pentium, but I doubt that debian will run on it at all, even when
> > ignoring the fact that the thing also only has 96M of ram, which is
> > to small to load a ramdisk and debian only targets i686. So the bug
> > might only apply to a very specific processor, unless there is a
> > patch
> > in the debian package.
>
> There are no relevant patches in the Debian package. This is exactly
> the problem with supporting old and obsolete architectures, that are
> very difficult to find in the wild: things break in weird and
> incomprehensible ways, and nobody is able to fix them. This is one of
> the main jobs of porters: if you can't triage and fix this issue, then
> it's likely you won't be able to triage and fix other architecture-
> specific issues either, as this is very very likely a hidden compiler
> toolchain issue. The effort required to have a release architecture
> officially supported in Debian goes way beyond "I have an old machine
> under the desk and can build some trivial packages", I am afraid.
>
> -- 
> Kind regards,
> Luca Boccassi
>
Looking at the upstream bug report, the issue seemed to also affect
theĀ  Intel J1900, which is a amd64 machine, so the bug is architecture-
independent. I'll look if I can get either of those processors, though I'd
doubt that.

However I do get the point that toolchain issues are a major problem and
can be a blocker for supporting an architecture, having encountered some
bugs when porting globulation2 to Haiku.
I'll look at the debian build tracker to have a look at broken packages.
Currently though I still don't really understand how the tracker works,
I'll first try to get a grasp about that.

Kind regards,
Maite Gamper

Back to linux.debian.devel | Previous | Next | Find similar


Thread

Re: About i386 support Maite Gamper <victor@wenzeslaus.de> - 2024-06-14 13:20 +0200

csiph-web