Re: remaining packages not rebuilt since May 2012
Svante Signell, le Fri 05 Jul 2013 15:37:55 +0200, a écrit :
> On Fri, 2013-07-05 at 12:33 +0200, Samuel Thibault wrote:
> > Svante Signell, le Fri 05 Jul 2013 12:09:43 +0200, a écrit :
> > > On Fri, 2013-07-05 at 11:39 +0200, Samuel Thibault wrote:
> > > > Thanks for the reminder, even if I had already taken care of checking
> > > > those recently :)
> > > >
> > > > Ansgar Burchardt, le Fri 05 Jul 2013 11:15:23 +0200, a écrit :
> > > > > unstable | atlas | 3.8.3-29
> > > >
> > > > There's apparently a computation difference here. The version currently
> > > > in experimental does build fine, however.
> > >
> > > I think the current version in unstable is 3.8.4-9.1, not 3.8.3-29?
> > That's precisely his point: hurd-i386 is out of date.
> > > There was a patch sent to debian-hurd for review in September 2012.
> > > http://lists.debian.org/debian-hurd/2012/09/msg00009.html
> > >
> > > The second part of that patch still applies. (And would add GNU as a
> > > unique architecture for the experimental version too)
> I can limit the patch to make 3.8.4-9.1 build without introducing a GNU
> architecture. However, maybe this is not necessary, since the
> experimental version would reasonably move to unstable soon.
That is what I meant.
> > Well, since the Linux port seems to be working in experimental, I'd
> > rather not introduce a GNU variant, which would inevitably lay back
> > within some time.
> Why would it lay back in time if adopted upstream?
Because we have seen that happen in various upstream projects. See
xosview for instance: there was zero porting needed actually. Just
making trivial Makefile changes exactly like for other OSes...