FESCo Meeting Summary for 2008-03-20

[Patrice Dumas]

On Fri, Mar 21, 2008 at 08:46:33AM +0000, Kevin Kofler wrote:
> Patrice Dumas <pertusus <at> free.fr> writes:
> > I don't really get it. There are many people in favor of blocking
> > packages until a guideline is done (at least I am, until it becomes
> > easier to force a packager to fix his package post import), they will 
> > report to FESCO when the packaging of a kind of packages is a mess (as 
> > it was for java packages).
> 
> If they notice... Not all technologies are as prominent as Java.

Indeed. But it is always like that, only people who care can block and
yiel. That's why I try to do merge reviews, to avoid lightly done
reviews for packages where packagers don't really care about quality
that are nevertheless important, but, of course this means that I have
to look at the merge review and at the package...

> > Part of fedora objective is also best quality in packaging. I don't know
> > if it is written somewhere, though.
> 
> But there's a fine line to walk there: having a perfectly-packaged kernel 1.0 
> is completely useless. We can't sacrifice being current entirely in the name of 
> packaging quality.

Sure. But the issue is not really this one, ther are always trade-offs.
The issue is more with packagers who don't care at all with packaging
quality and can only be coerced to care at review time. It is not
about searching every time the best quality, but more about avoiding
packages of very low quality. I have seen more than once people doing
'blitz' review/submission of package of very low quality. If there are
way to stop such packages when it is done on a large scale and do
guidelines to force submitters and reviewers to care at least a bit
about quality, then it is better.

--
Pat