Package EVR problems in Fedora 2007-07-27

[Michael Schwendt]

On Fri, 27 Jul 2007 23:06:37 +0900, Mamoru Tasaka wrote:

> Michael Schwendt wrote, at 07/27/2007 10:42 PM +9:00:
> > On Fri, 27 Jul 2007 14:24:12 +0100, Daniel P. Berrange wrote:
> > 
> >> On Fri, Jul 27, 2007 at 08:06:27AM -0400, buildsys fedoraproject org wrote:
> >>> berrange AT redhat.com:
> >>>     python-virtinst
> >>>       F7-updates-testing > F8 (0:0.200.0-2.fc7 > 0:0.200.0-1.fc8)
> >> This report is bogus. I built the F7 and F8 versions within minutes
> >> of each other yesterday mid-afternoon. The F8 version is 0:0.200.0-2.fc8
> > 
> > Your package is not in the rawhide repodata, and it is not tagged f8-test1
> > either. That's all that counts.
> 
> Well, I don't know how this EVR problem report is created, however

It's a script in fedora cvs /fedora/upgradecheck which I've run manually
recently after we had run it automatically after every Extras push.

> is it possible that we compare F7 repodata and koji repodata, i.e.
> the repodata on
> http://koji.fedoraproject.org/static-repos/dist-f8-build-current/i386/repodata/
> while rawhide is frozen until F8T1?

Yes, good suggestion. Thanks for that. 

But meanwhile I've been told some people in IRC spread false rumours about
the results of this script. Like claiming that the report would be useless
as it would only report false positives because the FE release procedure
doesn't push packages through updates-testing like bodhi does. *urgh* That
is a false claim. Lots of issues in the report are unfixed for many weeks.
Another packager told me he ignores the private reports because he
boycotts bodhi until there will be a "make release" or similar. With
opposition like that and spreading of false rumours, I rather stop running
the script altogether. It has become another thankless effort with no
backup from FESCo.