Re: Upcoming Debian multiarch support (amd64, sparc64, s390x, mips64) [affects sarge slightly]

[Goswin von Brederlow <brederlo@informatik.uni-tuebingen.de>]

Steve Langasek <vorlon@netexpress.net> writes:

> On Mon, Jan 12, 2004 at 07:36:01AM +0100, Goswin von Brederlow wrote:
> > Steve Langasek <vorlon@netexpress.net> writes:
> > 
> > > On Sun, Jan 11, 2004 at 08:00:36AM +0100, Goswin von Brederlow wrote:
> > > > after fiddling around with the multiarch support for amd64 for some
> > > > time now it looks like our developement plans and ideas for amd64 (and
> > > > hopefully the rest too) have stabelized. Since this involves several
> > > > changes to key components I would like to give everyone the heads up
> > > > about what we try to do and how we plan to do it. One minor change to
> > > > the /var/lib/dpkg/status file I would like to get implemented and
> > > > included in sarge. 
> > > 
> > > Regardless of the merits (or demerits) of this proposal, I object to
> > > trying to implement any changes to /var/lib/dpkg/status before the sarge
> > > release.  We are supposed to be in the last stages of the release cycle,
> > > and this is not the time to be trying to sneak changes to such a
> > > fundamental piece of our package system into the release.
> 
> > The change is ~10 line to dpkg with no changes to other software or
> > breaking compatibility.
> 
> Sorry, but that's always the assessment of a change's proponent, and is
> therefore not persuasive.  The kind of thorough testing that's called
> for when making a change to such a basic part of the system would almost
> certainly push back the release date.

After talking with the doogie on irc its aparently just a one line
change because its allready planed for in preparation of coming
multiarch support but there is an explicit line removing it for now.
Killing that line is aparently all it takes.

> > Are we realy in the last stages of the release cycle?
> 
> In spite of general attempts to continue pushing new and untested code
> into unstable, yes.  Steady progress has been made on getting red-letter
> packages from unstable into testing, which among other things means that
> there are few excuses left for those packages that *haven't* gotten
> their RC bugs fixed in sarge.  Getting rid of the last 350 RC bugs will
> be easy, when they're all package removals.

I'm more concerned about the debian-installer for non mainstream
archs.  A lot of ports have still a very limited support at best.

> > If the change is not added to sarge sarge+1 will have to deal with the
> > problem. Its not too much of a problem, since an upgrade from sarge
> > i386 to sarge+1 amd64 in a single step won't be possible anyway but it
> > would remove one big obstacle. The problem is that the status file
> > can't be changed in a preinst or postinst script since dpkg would
> > overwrite any changes when the script is done.
> 
> I have no objections to the idea of adding proper multiarch support to
> Debian, but your proposal comes way too late for these changes to be
> reasonably considered for sarge.

The only thing aimed for sarge is keeping the "Architecture: <arch>"
line in the status file. Thats a must for any multiarch support thats
going to come and its way easier to start multiarch support with the
line already present. Maybe that wasn't to clear.

Anything else for the multiarch support is never going to be stable
enough for sarge.

MfG
        Goswin