Re: "testing" improvements
En réponse à John Goerzen <jgoerzen@complete.org>:
> On Thu, Feb 27, 2003 at 04:48:28PM +0100, Jérôme Marant wrote:
> > Force packages into testing? You must have missed the point.
> > I said let packages into testing for architectures where nothing
> > prevent them to. Why would mipsel failures block x86 packages,
> > for instance?
>
> Sometimes bugs manifest themselves as compile-time failures on one arch
> and
> run-time failures on another.
Most of the times, yes.
> But in any case, the bug is in the source package, so why should
> binaries
> generated from a buggy source be allowed in?
A buggy source does not necessarily generate a buggy binary.
> Also, remember that testing is the foundation for the next stable.
> Imagine
> the support nightmare of having 12 different versions of the same
> package in
> stable. The work for the security team would be atrocious.
Well, you are obviously right. But, we can see many people
either using unstable or using unstable backports on stable.
It doesn't make sense.
But in fine, fewer do use testing: I personaly would like
them to use testing instead because testing is a pre-stable
release.
I didn't say that we cannot wait arches to be synced at freeze
time, but working packages on some architectures could enter
testing quickly, encouraging people to use testing and providing
bug reports quickly. Hence, people could test while we are fixing
bugs on problematic arches.
Cheers,
--
Jérôme Marant <jerome@marant.org>
<jerome.marant@free.fr>
http://marant.org
Reply to: