What Fedora makes sucking for me - or why I am NOT Fedora

[Jeff Spaleta]

On Wed, Dec 10, 2008 at 7:49 AM, David Malcolm <dmalcolm at redhat.com> wrote:
> Rationale: it's my belief that a rebase to a different upstream version
> typically carries greater risk of destabilization than targeted
> patching.  Thus, a bump to a more recent upstream should receive more
> testing than a packaging/patch change.

hmmm......I like that.  However that would still require having people
consume updates-testing and reporting in bodhi.   Do we have enough
people doing that? Do we have a good picture of how many are consuming
testing right now?  If we got a ratio of the number of ips in the
mirrormanager logs for updates-testing to updates-stable in say the
last week of F9 mirrormanager activity we'd have a baseline
understanding of the percentage of the base which is consuming
testing.

I'm going to talk about my experience for a second, mainly because I
like talking about me.  For most of the packages I maintain I don't
think I've ever seen karma raised to 3 or down to -3 for any package
I've had in testing...even when the update was pushed specifically as
a bugfix. I'll maybe get the reporter to chime in, but maybe not in
bodhi it maybe in the bugticket as a new comment (and it maybe by
consuming the koji builds directly before the bodhi push).  I'm not
sure I'd ever see an update for one of my packages pushed t stable if
I waited for bodhi karma to pile up, my packages just aren't that
important and don't appear in any default packageset for a spin. Maybe
that's different for more critical or more important packages.

-jef