[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Debian subprojects list needed



On Sun, Feb 24, 2002 at 01:39:03PM +0900, Junichi Uekawa wrote:
> Jeroen Dekkers <jeroen@dekkers.cx> cum veritate scripsit:
> 
> > On Sun, Feb 24, 2002 at 11:17:39AM +0900, Junichi Uekawa wrote:
> > >   - Debian on Athlon Project (!) :P
> > 
> > This doesn't have to be a seperate project if we just implement
> > http://master.debian.org/~brinkmd/arch-handling.txt after 3 years of
> > ignoring it.
> 
> No, we don't need proposals without solid working 
> test system.
> 
> The purpose of my project is to create a testbed, and 
> test implementation so that the system can be evaluated.
> 
> Proposals without solid testable implementation is 
> just useless. (er... actually, unusable :P )

If somebody volunteered to work out the details and make test
implementations I don't see any problem. However, if nobody replies it
and offers a little help (not everything can be done without help of
some dpkg hacker, for example), busy people who volunteered tend to
work on other tasks they want to do. I'm not going to blame anybody on
this, but I don't think it's fair that the Hurd port is blamed for not
wanting intergration with Debian, which was said so on IRC
yesterday. But I'm volunteering to do some work. I'm certainly
interesting in the Debian on Athlon Project and how they do
everything.

The Hurd port was also blamed for doing everything different. Now we
don't do _everything_ different, but we do some things
different. Things we consider better the way we are doing, but we try
to be compatible with the rest of Debian. But please face it: the Hurd
port is different from all the other ports in Debian. It's a whole new
OS, not just a port of GNU/Linux to another architecture. Also, the
Hurd itself is not just another implementation of the unix
kernel. Altough we do things different, we try to be compatible, with
the POSIX standard for example. If the are bugs in the compatibility,
please let us now about it. We can't fix things without bug reports.

For the FHS: we care about it, especially as it is in the Debian
policy. However I don't expect the FHS to include all things possible
on the Hurd. The Hurd has a very flexible filesystem namespace, I
could write a document longer than the FHS about all his
possibilities. ;-)

We are compatible however, we have the /usr -> / symlink, and *all*
packages should reference files in the /usr location, it doesn't
matter that's physically in another place. Making /usr a directory
should also work and report if it doesn't. We have our reasons to have
this symlink at the moment. We want to use shadowfs in the future,
which in our opinion solves the problem of needing a small partition
to boot from in a better way.

For the autobuilder issue, turle is in a long term hibernation. The
problems we had with glibc are fixed now and we are going to start
building packages linked against our new libc0.3 (Yeah, also different
from GNU/Linux, we don't consider our ABI stable enough yet) using the
normal autobuilder. Just wait a few weeks and the you can see the
number of hurd-i386 packages increasing. And you can see bug reports
saying that your package doesn't build of course. But you don't have
to wait for that, you can fix your MAXPATHLEN problems already. :))

For the people complaining about the intergration of the Hurd port: We
have a mailinglist to ask questions on and file bugs (but RTFM please
first). AFAIK the people saying on IRC that we don't want intergration
never asked anything about the problems they saw. If you read our
archive, you see that we have talked about the FHS. We care and I'm
going to write a proposal for the OS-specific annex and check if we
are compliant with the FHS.

Now let's stop blaming others. Let's think about the future instead of
the past. We are also getting the *BSD (* includes GNU/) port(s) now,
let's fix the problems the Hurd and *BSD ports are having. Let's
cooperate. We all have a common goals, only trying to do it in
different way: Making Debian better, making Debian *the*
distribution for everything and supporting Free Software. Wouldn't it
be nice to have total world liberation? ;-)

I don't think one kernel fits for all purposes. Of course I don't deny
I think the Hurd port will fit for all purposes in the future, but the
people working on other ports also think that about their ports. That
doesn't have to be a problem, Debian can have different
kernels. Ket's just hack and see what the future brings us. Maybe we
win, maybe you win, maybe we both win and Debian will have 10
different kernels in the future. Wouldn't it be cool to have the
Debian installer question: "What kernel do you want to install?" I
think that's easier to do than it looks like.

We don't have to fight each other, we could cooperate instead.

Jeroen Dekkers
-- 
Jabber supporter - http://www.jabber.org Jabber ID: jdekkers@jabber.org
Debian GNU supporter - http://www.debian.org http://www.gnu.org
IRC: jeroen@openprojects

Attachment: pgpIWki8qeGnj.pgp
Description: PGP signature


Reply to: