On Tue, May 23, 2006 at 11:30:57PM +0200, Jan de Groot wrote:
> On Tue, 2006-05-23 at 23:10 +0200, Tobias Powalowski wrote:
> > Hi
> > hmm i don't know how this was all organized but i see some real problems in
> > x64 organisation:
> > - cvs get broken by ppl that don'T even inform us about their doing
> > - they should be at least on the dev team that we can talk to them and they
> > can inform us about their doing.
> > - me is scared by those ppl that say on forum they want to be more bleeding
> > edge then i686, when all packages should be the same status not that one
> > port runs in front of the other.
> > - there was no discussion about that at all on ML how to handle that stuff.
> > - the new gcc,glibc stuff rebuild for 0.8 will be a big task so this must be
> > organised well, else it becomes a fiasco.
> > so please hear that call as it is now it causes a lot of trouble.
> > any other opinions out there?
> One more thing: the amd64 port comes without multilib support. lib ->
> lib64 symlinks are all over the place, etc. IMHO we should have some
> quality standards for this:
> - kernel should support 32bit binaries (AFAIK it does)
> - all libs should be installed in either /lib64 or /lib, not both.
> - 32bit runtime libraries and crosscompilers should be available as
> separate packages. Libs should go in /lib, or if this is chosen for
> 64bit already, /lib32.
> Packages should be in sync with i686, not ahead of it. One thing that
> doesn't look so good: why did we start this project with x86_64 in the
> first place, the ppc port is much more grown up. If ppc would have
> worked out, we could have included x86_64 later then.
While I love being a go-between from the x86_64 community and the devs, we
should probably take these discussions over to the arch-ports list so that
everyone can comment on them.
I'm CC'ing that list with this email.
If you understand, things are just as they are. If you do not understand,
things are just as they are.