Re: Software version issue
On Tue, 7 Oct 2014 23:29:34 +1300
Chris Bannister <cbannister@slingshot.co.nz> wrote:
> On Tue, Oct 07, 2014 at 09:05:08AM +0100, Joe wrote:
> > On Tue, 7 Oct 2014 09:52:51 +0300
> > Andrei POPESCU <andreimpopescu@gmail.com> wrote:
> >
> > > On Lu, 06 oct 14, 21:26:33, Joe wrote:
> > > >
> > > > So your choice is between testing and unstable, Jessie and Sid.
> > > > Sid is a rougher ride, more likely to have things broken at any
> > > > time and without warning, but for the next (approximately) six
> > > > months it is the only distribution which will receive new
> > > > software. How new do you need?
> > >
> > > Actually, during the freeze the Release Team asks all Maintainers
> > > to upload new stuff to experimental, to keep the path unstable ->
> > > testing for RC bug fixing as much as possible.
> > >
> >
> > Yes, I know there is a major upheaval in Sid after release, but I
> > don't think I've ever noticed Sid being frozen...
> >
> > What does make sense is not to allow Sid to get too far ahead of the
> > frozen testing, at least in terms of major system architecture.
> > Testing must remain upgradable to the latest state of Sid without
> > too many breakages, once the freeze is over.
>
> I don't know about that, after all, the main raisin d'etre of Debian
> is the stable release. Things are normally a bit erratic after a
> stable release. It's a good idea to wait for things to cool down a
> bit. This is the time when you either feel out of your depth running
> testing or Sid or you cruise along riding the bumps and troughs
> knowing when to ease off on the throttle and when to pour on the
> juice. We will soon find out the ones who have grabbed the controls
> and just put it on full throttle and hope for the best. :)
>
What else can you do? That is what Sid is for, to find the broken bits,
and if you're picking your way cautiously among the potholes, you're
not finding out how deep they are. My main workstation is Sid, I
upgrade every upgradeable package pretty well every day. If something
breaks, I look into reporting it, though even then I'm rarely the first
to find anything.
But you're right, it's all about stable, and by also maintaining a
continuously more-or-less operational distribution along with an
equally continuous test-to-destruction distribution, you're helping to
ensure that *every* stable can be upgraded to the next. That's the USP
of Debian, that a stable upgrade is *always* possible. Unstable and
testing show how to get from A to B, because they go ahead and do it,
without missing a beat. Mostly.
--
Joe
Reply to: