[Gnumed-devel] Re: Unstable vs testing

[Andreas Tille]

On Sat, 28 Feb 2009, Karsten Hilbert wrote:

> > Since gnumed has many dependencies, is it better to *not* be
> > installing gnumed-client and all of its dependencies from unstable?
> When you run testing all packages from unstable are slated for
> moving to testing within 10 days anyways (barring a blocker).

Additional comment: There are people out there who are running
unstable and there is even a distribution (Sidux) which claims
unstable is good enough.  I personally think *unstable* has its name
for a reason and so I avoid it "wherever possible".  It is finally
your decision based on your knowledge whether it is possible
for *you* to avoid unstable.

> > If the package usually moves from unstable into testing within about
> > 10 days then I wonder whether the needs of people who would depend on
> > gnumed for real use,

I would *never* trust people who depend on a GNUmed version which
is sitting in unstable.  If there is a burning (=security) problem
it should be delivered via the security update mechanism in stable.
If it is "just another new version" people should not really
*depend* from it.

> > It seems like we are normally testing release candidates via
> > tarballs. This makes we wonder whether, after a gnumed-client package
> > is built and is moved into unstable, do we ever get RC bugs filed
> > against the gnumed package in unstable?
> We do at times.

... as any piece of software.

Kind regards

        Andreas.

--
http://fam-tille.de