Re: Many packages missing from testing
On Saturday 12 November 2005 01:21, Johan Kullstam wrote:
> loos <loos@qt1.iq.usp.br> writes:
>
> [snip rant against "testing"]
>
> > I just totally agree with you. A little difference, I switch my
> > production machines (stable) to testing somewhere during the "frozen"
> > time (of course using testing real name. I prefer having a manual
> > control on the oldstable->newstable update. I am around since ham and
> > this worked without problems for me.
>
> I agree. The fixed names are much better. There was a thread here a
> while back (6 months, a year?) about making the default be a fixed
> name like "woody", "sarge" or "etch", rather than "stable". I think
> that would be a much better default.
>
> > My desktops use unstable.
>
> As are mine. Sid is pretty solid for me so far. I can recover from
> most of the mishaps. But at work I do worry since I could lose hard
> if things go really badly. I guess that's why they make stable. But
> that's so boring ;-).
>
> > The problem is always the same: Newbies don't understand the sense of
> > the word "unstable" as used by Debian.
> > In fact they lack understanding what a distribution is, and therefore
> > what a stable (or unstable) distribution is.
>
> Exactly. I was using "testing" for a while and got tired of losing
> when a package broke and wouldn't get fixed for ages.
>
> Of course, a savvy user could default to testing and drag in unstable
> (with whatever pre-reqs) whenever a breakage occured. Perhaps this
> method could be made more known.
>
Wouldn't pinning work very well in this case to allow a mixed testing/unstable
system? The trouble packages can then be installed from unstable using the
-t option, with the majority of the rest of the system runs at a testing
level (for example all the non GUI stuff).
Chris
--
PD Dr.-Ing. C. Hurschler
Bodenstedtstr. 13
30173 Hannover
0172-5940909
Reply to: