Re: Frozen distribution?
On Mon, Feb 19, 2001 at 12:23:04AM +1000, Anthony Towns wrote:
> On Sun, Feb 18, 2001 at 12:57:53PM +0000, Julian Gilbey wrote:
> > Why suddenly change the model like this?
>
> The model's already changed: we went from
>
> stable
> frozen (temporary)
> unstable
>
> to
>
> stable
> testing (permanent)
> unstable
Yes, but neither stable nor unstable changed their meanings. testing
was a new concept.
> > Would the following not be
> > better and perhaps less confusing, still using a four-tier setup:
> >
> > stable frozen testing unstable
> > [...]
>
> This is somewhat confused.
I agree :-)
> "testing" is maintained by a script, it is not uploaded to directly. What
> you describe would be mostly sensibly named:
>
> stable testing frozen unstable
>
> where stable is updated as it is now, testing is updated from frozen
> instead of unstable by scripts, and both frozen and unstable can be
> uploaded to directly. At release, frozen is removed entirely, testing
> becomes the new stable, and is also forked into a new testing which is
> once again updated from unstable.
Yes, that sounds good.
> The problem with doing something like this is what I described before:
> bugfix uploads need to go to both frozen & unstable, which causes problems
> for the autobuilders.
That is a problem we should be fixing, surely, not working around.
> It also introduces a temporary distribution, which is a nusiance to
> maintain.
Any more so than the old frozen?
> And in truth, it doesn't seem much of a win over just using
>
> stable testing unstable experimental
>
> in those roles. The only benefit it gets is that the few people who fork
> their packages during the freeze don't have to file a bug report to get
> their package moved into unstable from experimental (and as far as the
> ftpmasters are concerned, that's a fairly trivial bug to resolve with
> "heidi").
I question whether only a few people fork. I simply don't know any
stats.
What I do predict is the following:
As all of the tools currently send packages into unstable by default,
you are going to find that (almost) everyone will continue to send
things into unstable after the freeze. And then unstable will
continue to be just that: unstable. It will be really hard to
separate things which should be go into frozen from those which
shouldn't, and we'll probably end up with the messy situation of
library incompatibilities and such like.
I really think that changing the meaning of unstable in such a
fundamental way is a really bad idea.
> > (b) People believe that the next freeze is not going to be another 18
> > months away.
>
> The last freeze began thirteen months ago, I don't see why you'd be
> worried that the next one would be further away.
It's an issue of perception, not reality.
Julian
--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Julian Gilbey, Dept of Maths, Queen Mary, Univ. of London
Debian GNU/Linux Developer, see http://people.debian.org/~jdg
Donate free food to the world's hungry: see http://www.thehungersite.com/
Reply to: