Re: [Fwd: Re: Debian based GNU/Solaris: pilot program]
On Wed, Nov 02, 2005 at 06:31:00PM -0800, Erast Benson wrote:
> On Thu, 2005-11-03 at 01:14 +0000, Matthew Garrett wrote:
> > Alex Ross <firstname.lastname@example.org> wrote:
> > > Michael Banck wrote:
> > >> If so, do you plan to use Debian's mailing lists and bug
> > >> tracking system for development?
> > >
> > > No. We have ours: svn, Trac, and mailing lists.
> > It's unlikely that you'll be accepted as an official Debian port unless
> > you're willing to use the Debian bug tracking system. It's not
> > reasonable to expect Debian maintainers to be willing to copy bugs to a
> > completely different bug tracking system in cases where it turns out to
> > be a Solaris-specific issue.
> on another hand, is Debian community willing to be not just GNU/Linux
> centric and put some work on GNU/Solaris too? If yes, we could
We do have non-Linux ports in the works (in various states of completion).
Typically they don't get released because there is insufficient interest to
get them to the quality level needed for a stable release. This lack of
interest probably stems from a "Linux is OK for me" viewpoint rather than an
"all these non-Linux ports are useless" opinion -- that is, apathy rather
A released Debian/Solaris would, in all likelihood, enhance Debian in all
sorts of ways, like porting a regular program to 64-bit and big-endian
architectures cleans things up.
> on another hand, Ubuntu has its own tracking system, so GNU/Solaris is
> not the first one. Even though Ubuntu is GNU/Linux system...
It's GNU/Linux, but not Debian. It's a derivative. The question here isn't
whether you want to use some Debian-derived technologies in your port (which
you're free to do with or without any input or cooperation with Debian
itself) but whether you want to be part of A Debian Release.
> on another hand, GNU/Solaris uses different kernel and libc, which
> brings many non-Debian-related issues into play.
Yeeeeehah! As I recall, there were plans to produce a non-glibc port of one
of the BSDs, so there's precedent at some level. Being
not-so-glibc-dependent would also benefit projects like the guys trying to
rebuild Debian for uclibc (or one of the other itty-bitty-libcs) for use in
the embedded space.