Delays in package processing

[Michael Schwendt]

On Fri, 21 Dec 2007 10:12:43 +0100, Thorsten Leemhuis wrote:

> On 21.12.2007 09:46, Michael Schwendt wrote:
> > On Thu, 20 Dec 2007 10:19:48 -0900, Jeff Spaleta wrote:
> > [...] 
> > *I* believe we flood our users with too many rushed/untested updates. It
> > feels more and more like a rolling release which is not too far away from
> > Rawhide.
> 
> I think a kind of rolling release is something good -- especially as
> hardware-support in Linux is not done by separate drivers (like on
> Windows) and instead often bound to packages (like the linux kernel,
> sane, or hpijs). Thus is we IMHO want new version of those packages in
> the repo to support new hardware instead of forcing them to use rawhide
                                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^
> or waiting up to six month until the next release.
     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Hyperbole. There may be feature additions in updates, provided that they
are subject to a reasonable amount of testing.

> But On the other hand I agree that there are many rushed/untested updates.

All I wish is that we don't throw away the results of the development
cycle so carelessly: our choice of what components we add together to
build packages, the testing, the freeze. Some packagers complain about
bureaucracy, such as having to mail rel-eng when rawhide is frozen and
they want a rebuild to be added. As soon as the distribution is released,
the flood-gates open and packagers pipe out lots of version upgrades in
form of updates which are pushed quickly. And the stream doesn't stop,
because next week another "minor" version update is offered, again bearing
the risk that it breaks. Then, the only excuse is that Fedora is not
suitable in production environments. Do we want to drive away users to
other platforms?