[Mageia-dev] Opening backports (was Re: [Mageia-sysadm] Using SQL database for youri)

[Samuel Verschelde]

Le jeudi 29 septembre 2011 21:10:42, Samuel Verschelde a écrit :
> Le jeudi 29 septembre 2011 20:55:43, Florian Hubold a écrit :
> > Am 29.09.2011 14:13, schrieb Olav Vitters:
> > > On Thu, Sep 29, 2011 at 10:24:56AM +0200, Samuel Verschelde wrote:
> > >> backports_testing, but it means that we would lose the current
> > >> checking that updates come from the 1/updates branch)
> > > 
> > > Long term I prefer more checks because I do make mistakes. But maybe
> > > lax as a temporary ("a few weeks") solution until it uses the SQL
> > > database?
> > 
> > Well, is it really that necessary for a matter of some weeks to speed up
> > backports opening? Shouldn't we care first that sustainability of Mageia
> > is ensured first?
> > 
> > It's not like backports candidates will lessen workload on QA, rather the
> > opposite. That's my personal opinion.
> 
> Some genuine user needs can only be covered by backports per our policy, I
> don't see why we should wait more (already 4 months since the release of
> Mageia 1). QA team already said it's ready to test backports, and if we
> have difficulty we simply will find means to get more testers. Also,
> updates will still remain the priority.
> 
> Opening backports media does not mean that we will start backporting
> everything. But a policy has been chosen, after a loooooooooong debate, and
> there seems to be a simple way to open those media to package submitting.
> 
> Also, the sooner we have backports, the less there will be external third-
> party repos with all the problems (upgrade, support) that causes. There
> already are, don't let them too much space and rather invite their
> packagers to backport *inside* (as long as they stay within the policy of
> course).
> 
> About Mageia's sustainability, we already are working to that (QA team is
> operational, package adoption campaign is undergoing...), and having
> backports media won't prevent us from continuing.
> 
> Also, the "few weeks" is a really vague estimation, my guess is it risks to
> be rather months because sysadmins seem to have a big TODO list a little
> resources. That's why I'm proposing 2 "technically simple" ways to move
> forward.
> 
> Best regards
> 
> Samuel Verschelde

If needed, the tremendous amount of packages in Blogdrake's 3rd party media 
shows how much backports are needed by users, whatever we as packagers can 
think of it.

I'm still convinced that opening backports right now, using one of my 2 
proposals (if I haven't overlooked a technical difficulty), would be an 
important step forward for us.

Best regards

Samuel