Don't put new packages through updates-testing

Michael Schwendt mschwendt.tmp0701.nospam at arcor.de
Sat Jun 2 11:28:36 UTC 2007


On Sat, 02 Jun 2007 16:16:27 +0530, Rahul Sundaram wrote:

> Michael Schwendt wrote:
> > On Sat, 02 Jun 2007 15:25:04 +0530, Rahul Sundaram wrote:
> > 
> >>>>>> The base functionality must still be tested.
> >>>>> That is much too vague.
> >>>> We need to sit down and define the details more precisely if we agree 
> >>>> with the fundamental idea.
> >>> Don't you fear that it would drive away reviewers?
> >> I does raise the bar a bit on the reviewers. I fear more that we are 
> >> driving away users by the mentality around pushing packaging and waiting 
> >> for bug reports to arrive that the package is basically broken.  If we 
> >> continue along that path more we won't be left with much users.
> > 
> > You're pushing at the wrong side.
> > 
> > That FESCO has neglected to pursue some important goals [of the past] is
> > no secret. 
> 
> I am not sure assigning blame is going to be useful in changing 
> anything.  It is a fact that we could do better in our QA processes 
> without directly involving FESCo.

Unfortunately, you won't find enough volunteers who would be willing to
jump over unnecessary hurdles before realising that they would likely not
get the proper privileges to clean up wreckage in the distribution. Not
even sponsors have the privileges anymore to intervene in CVS or buildsys
(thanks god the people I sponsor are the FESCo chair, FESCo members and
pleasant packagers who don't cause any trouble anymore). FESCo is directly
responsible for implementing stuff that makes many things unnecessarily
harder. AWOL procedures, ACLs, after years there still is no team that
could step in and apply emergency fixes. Just like FC6 updates still
contain broken deps, there is no indication that F7 will be any
better. Already in the first week, there are broken dependencies which
also make yum upgrades from FC6 much more difficult if not impossible. And
those that would be fixed quickly, are not fixed because they sit in
updates-testing or because packagers need to take time and learn about
bodhi'n'stuff.

You call it "blame", others call it "opening eyes". Historically, it has
been necessary several times, too.




More information about the fedora-devel-list mailing list