Re: Granual release proposal
On Fri, May 24, 2002 at 11:49:47AM +0300, Richard Braakman wrote:
> On Thu, May 23, 2002 at 03:30:52PM -0500, Andrew Lenharth wrote:
> > Problems seeking solutions:
> > 1. The whole distribution must stablize at once.
> > 2. Freeze time caused by 1
>
> The "caused by 1" part may have been true for potato, but it doesn't
> seem to have been true for woody.
Sure it does. Did we not just have to stablize every package to be included
in the release? Get them all ready for a simultanious release?
Sure, we frooze in stages (policy, base, packages), but in the end,
we still had to get all the RC bugs out of every package so we could
release them all at once.
> > 3. new upstreams can take over a year to make it to stable, see 1
>
> Wouldn't that be "see 2"? If we could have a shorter release cycle,
> then this problem would be solved.
Yes. But as I don't beleive we can have a noticably shorter release cycle.
The archive is simply too big. Complex packages hold up simpler packages.
We don't have XFree 4.2 because it would delay the release too much. Other
major sets of packages in stable are very old because they can only be
updated on a release which happens once every other year (or so), yet they
are release quality.
> > 4. no modularity
>
> This sounds more like "modularity" is a solution seeking a problem :-)
I would say 4 is the cause of 1-3. Further, I claim it has 2 soluitions:
The one you mention (modularity) or huge amounts of man power applied in
a more concentraited time period. As we have no control over the
application of man power, I simply say that we should increase modularity.
Andrew Lenharth
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: