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

Re: 2.4.0!



I figure I will chime in with my $1.25(from inflation).  First, I see that
Debian has the unfortunate tendency to put EVERYTHING together all at once, then
fix all the parts that break, then fight over dependancies.  This makes a lot of
people upset, and causes a delay in releasing the next version.  My solution,
which I feel makes the most sense, is to start with base.  Since base should not
have any dependancies on things NOT in base, that's where all work should start
for the next release.  While this work on the new base is being done, the
package maintainers can continue their work on their packages, which can be put
into a maintenance release of the previous version.  Once base is stable, and
without any critical bugs, the next set can be added.  This will include the
usual utilities, as well as the base XFree86.  Obviously, packages that depend
on X will need to wait until XFree 86 is stable before they get added.  Nice,
clean, keeps the sense of what's important clear.  It also means that if all you
want is a server(without the need for XFree 86), you can have your updated
version of Debian much sooner.  The final release may take the same ammount of
time, but by layering the approach that is taken, it will keep things a bit
cleaner, and avoid some of the arguments that have taken place.  There are
things that would have to be worked out, but I'm sure EVERYONE would be happy if
glibc didn't get updated while work is being done with XF86 based applications.

							David Bristel


On Thu, 25 May 2000, Joseph Carter wrote:

> Date: Thu, 25 May 2000 16:43:57 -0700
> From: Joseph Carter <knghtbrd@debian.org>
> To: Richard Braakman <dark@xs4all.nl>
> Cc: debian-devel@lists.debian.org
> Subject: Re: 2.4.0!
> Resent-Date: Thu, 25 May 2000 16:42:23 -0700
> Resent-From: debian-devel@lists.debian.org
> 
> On Fri, May 26, 2000 at 12:55:09AM +0200, Richard Braakman wrote:
> > The implication is that packages that are generally considered 'important'
> > need less testing.  I don't buy that.  Something like a new gnome or
> > a new XFree will need *more* testing than the average package, not less.
> > They are also highly complex packages.
> 
> I believe the point is, and FWIW I agree, that a few key packages are
> easier to stabilize and fold into an already stable system than trying to
> stabilize 4000+ packages all at once.  Surely your experience managing
> this release indicates that clearly to you.  There is simply an immense
> amount of coordination that is required to freeze and release unstable.
> 
> On the same token, it's relatively easy by comparison to take a 2.2 kernel
> once the bugs are worked out of it and add it to Slink.  Or a new Apache.
> Or a new Gnome.  Or any other package which a significant majority of
> Debian users and developers feel needs to be in the stable distribution to
> keep it current with the needs of its users.
> 
> 
> We cannot add these things to Potato now.  I do not now and never have
> disputed that.  We should, however, commit ourselves right this moment to
> adding these things to Potato after release when it is believed these
> packages are in the context of Potato, good to release.
> 
> Same principle as the package pool applies:  It's easier to make a few
> packages stable for release with everything else that is already stable
> than it is to hack everything and then call a halt to development while
> people work to stabilize the work that's been done.  How many developers
> are already running Woody, regardless of Potato's status?  I know I am.
> 
> -- 
> Joseph Carter <knghtbrd@debian.org>               GnuPG key 1024D/DCF9DAB3
> Debian GNU/Linux (http://www.debian.org/)         20F6 2261 F185 7A3E 79FC
> The QuakeForge Project (http://quakeforge.net/)   44F9 8FF7 D7A3 DCF9 DAB3
> 
> <Valkyrja> java, hon, sometimes I really want to smack you.
> <Knghtbrd> Valkyrja - he'd enjoy it too much
> <Reteo> Valkyrja: yah, go ahead and do it... beat java into cappuccino! :-)
> 
> 
> -- 
> To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
> 



Reply to: