Delays in package processing

[Michael Schwendt]

On Thu, 20 Dec 2007 07:12:49 +0100, Ralf Corsepius wrote:

> Isn't testing what is supposed to implement the "delay queue", which is
> what you seem to be asking for.

It is circumvented too easily. Packagers raise the karma for their own
packages. Release people don't reject quick pushes from testing to stable,
which are not security related or critical. Updates-testing is just not
popular enough due to the bad smell of burnt guinea-pigs. The packagers
see it as a just another hurdle on the road to getting package updates
published. The users don't like a never-ending stream of updates to
download when it doesn't fix their very own list of problems.

> > The updates
> > repository is continuously flooded with version upgrades, which move
> > farther away from the tested gold release of the distribution only to
> > break due to new bugs, which then require further updates.
>
> At the same time Fedora+updates is suffering from bugs not receiving
> fixes in reasonable time.

Is this because they sit in "pending" without being pushed for several
days? Or is this because the packages contain bugs which are not fixed or
which are fixed only in rawhide?

> To put it bluntly:
> * As a packager, I feel strangled by current release practice.

What do you want? Daily pushes? Would that suffice? Immediate [automatic]
pushes of security/critical fixes when the packager clicks a button?

> * As a user I am gradually feeling annoyed by seeing bugs not getting
> fixed.

Our users frequently see update announcements, but the tools don't find
these updates because only the master site has them. It is not obvious
that it takes 24 hours or days for mirrors to offer the updates.

> * If I were still a "low bandwidth user" I would quit Fedora now,
> because updates are being pushed in "big chunks" blocking internet
> access for hours once a week, instead of being fed with "small chunks"
> in shorteŕ intervals.

The problem is that too many updates don't fix real-world problems or
issues reported in bugzilla tickets.

They are ordinary minor version updates with no or questionable benefit
(such as uninteresting changes, huge autotools updates, dangerous
from-scratch-rewrites, stuff that invalidates the results of the fedora
development testing period, spec "fixes" for packages that built fine,
superfluous %config modifications which trigger .rpm{new,save} and annoy
users who actually use the software).