Viktor Rosenfeld wrote:
>
> Ross Burton wrote:
> >
> > On Thu, 2001-08-30 at 17:09, Robert L. Harris wrote:
> > >
> > >
> > > hmm, Ok, give me a list of stability:
> > >
> > > potato most stable
> > > sid (unstable) next most stable
> > > woody (testing) "least" stable
> > >
> > > Problem is manage ment doesn't understand "shortly after" they want
> > > an "average # of hours/days" etc.
> >
> > Wrong order, unstable is less stable than testing. Testing is packages
> > being tested for stable.
>
> I've found unstable to be of better use than testing. The reason is
> that even bugfixes need at least 10 days to go into testing, whereas in
> unstable they could be included the next day. I've been badly bitten on
> some occasions by a testing dist-upgrade that left my system broken.
>
> Granted, unstable is also broken sometimes, but I have found that the
> errors are much easier to fix than testing errors. And if I can't get
> it to work, it'll usually repair itself with the next dist-upgrade.
exactly. I tried testing and had few long-term pronblems (weeks)
because of various reasons... with unstable there are about the same
number of problems (VERY RARE) and the6y are fixed fast (usually when I
find out about the probem the fix is already coming). Judging by mails
in this thread my experience is not unique...
erik
