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

Re: hurd does NOT need /hurd

On Mon, May 20, 2002 at 04:16:03PM +0200, Jeroen Dekkers wrote:
> > Just for some perspective, the thing that's stopping it from release
> > is that it doesn't work (it's been a year since I last knew what was
> > wrong, but missing partitioning tools, firewalling tools, an installer,
> > and massive incompatible bin format changes come to mind; I've no idea
> > how accurate they are as criticisms today though). 

partitioning tools: done upstream, needs only to trickle down
                    (we need parted > 1.6)
firewall: I don't see how this is release critical at all.
installer: oh well, yeah, definitely.  debootstrap won't be too hard
  (although the lack of translator support in tar calls for a different hack
  than packaging the /dev directory into a tar file), but that is only a
  small part of the installer that we have a native implementation for
  years, so we know the issues.  I don't know about busybox, for example.
  Maybe Glenn can clue us in.

the glibc ABI change (if that is what is meant with massive incom...)?
Debian GNU/Linux had two of them after the first release, IIRC. 
We had one before the first release.  Looks like we are in better shape :)
It was a good thing that it happened.

On my list is: Make GRUB boot from CD.  is that still an issue

> I think the things that's stopping it from a release are the number of
> packages ported, Marcus could correct me here if I'm right.

Personally, I couldn't care less if we release with 74% or with 86%, as long
as all the important packages people rightfully expect to be there are
indeed there.  I can immediately think of record locking, shared mem, and
semaphores, as well as pthreads (which is a biggie and a showstopper, which
is really an unfortunate combination).

And I personally want to see the switch to oskit rather sooner than later,
which adds another lot of driver related issues.

Well, it would be nice to release without a makedev package in the archive,
but the Hurd package control field seems to fix this at least for makedev
quit nicely.

Not everything of this is critical.  But I guess most of it could be done
for woody+1, if people are found to do the work.


`Rhubarb is no Egyptian god.' Debian http://www.debian.org brinkmd@debian.org
Marcus Brinkmann              GNU    http://www.gnu.org    marcus@gnu.org

To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Reply to: