QA-process for new packages

Michael Tiemann tiemann at redhat.com
Thu Sep 30 18:47:25 UTC 2004


Silke, I'll volunteer to be one of your reviewers.

M

On Thu, 2004-09-30 at 14:39, Michael Schwendt wrote:
> On Thu, 30 Sep 2004 19:26:37 +0200, Silke Reimer wrote:
> 
> > Some time ago I submitted a few packages on fedora.us. One of them
> > (gdal, Bug #1964) got lots of comments so I rebuilt the package and
> > announced it today. Due to several reasons it took me some time to
> > rebuild the package and meanwhile I have been set as owner of the
> > bug (and of all my other bugs (#1965, #2000 and #2001) as well).
> 
> The assignment of package request tickets to package owners has been
> announced and explained on this list around two weeks ago.
>  
> > Since I am not member of the QA team
> 
> Who said that? 
> 
> You _are_ a member of the QA team. The community does QA on packages
> in the queue. That's pre-release QA. Other members of the community do
> post-release QA and submit bug reports when they find something in
> the published packages.
> 
> > I don't really understand this
> > action. I thought that people that are new to fedora can submit
> > packages thus being submitter of a bug. Afterwards the QA team
> > assigns someone to do the quality assurance and the submitter will
> > have to fix the package if there are problems.
> 
> There's no such system. Doing QA on new packages and package updates
> is done by volunteers. And they are not assigned packages to QA, but
> they choose interesting submissions themselves. This system is flawed.
> Because if I reviewed and approved 200 different new packages, I would
> need to assure that any future update requests for those 200 packages
> are QA'ed, too, until the package developer reaches "trusted" status.
> 
> So, what I've been doing recently is to pick packages from new
> contributors and give them the chance to get a package published. In
> return, however, I'd like to see that they engage in QA and help other
> contributors. I've counted more than 60 different names in the queue,
> so theoretically, there are enough different people to choose from.
> Further, I monitor the REVIEWED queue, and I take a look at older
> package requests from the trusted developers, too, because they don't
> need QA for updates.
> 
> > So, my question is: Did I misunderstand the process? And what should
> > I do to ensure that a QA team member might look at my packages? Or
> > is this perhaps the normal process and I don't have to do anything
> > at all?
> > 
> > Thanks for all explanation (or some hints to any documention [1]),
> > 
> > 	Silke
> > 
> > [1] Yes, I already read: http://www.fedora.us/wiki/PackageSubmissionQAPolicy
> 
> The first paragraph here,
> 
>   http://www.fedora.us/wiki/PackageSubmissionQAPolicy#review
> 
> contains an important piece of information:
> 
>   [...] encourage other packagers to thoroughly check your package by
>   doing a good job in QA testing of their package. [...]
> 
> 
> Might be or might not be that all this changes when the merger with
> official Fedora Extras is complete. But again, according to the
> package submission and QA policy at fedora.us, it doesn't take much to
> get a package published...
> 
> -- 
> Fedora Core release 2.91 (FC3 Test 2) - Linux 2.6.8-1.541
> loadavg: 1.60 1.92 1.70




More information about the fedora-devel-list mailing list