On Wed, Aug 26, 2020 at 7:11 PM kevin <kevin(a)scrye.com> wrote:
On Wed, Aug 26, 2020 at 11:45:48AM +0200, Fabio Valentini wrote:
> On Tue, Aug 25, 2020 at 9:33 PM kevin <kevin(a)scrye.com> wrote:
> >
> > On Tue, Aug 25, 2020 at 08:14:38PM +0200, Fabio Valentini wrote:
> > > Do you think it would be a good idea to file bugs for those packages
> > > where a build for f33 was "forgotten"?
> > > Since I already have the "downgrade check" scripted [0] it's
a simple task of:
> > >
> > > - check if the package has attempted build for f33
> > > - if yes, check if it's FTBFS, and don't report a bug
> > > - if no build was attempted for f33, report a bug
> > >
> > > I see ~100 downgraded *binary* packages going from fully updated f32
> > > to fully updated f33.
> > > The number of affected *source* packages is naturally lower than that.
> > > The script also doesn't account for packages being renamed (but in
> > > those cases, proper Obsoletes and Provides need to be present during
> > > package review, so I don't think this is a problem).
> >
> > Yes, I think that would be lovely.
> >
> > kevin
>
> I'm on it.
>
> By the way, I'm finding some weird issues where packages are tagged
> with f33 in koji for a few days but an *older* build is in the repos,
> e.g. appliance-tools and ddgr.
I am not seeing that here. Could be a mirror issue?
I ran out check_latest_build script on f33 (which looks at each
package and tries to figure out if the last tagged one is the
'highest').
It gives:
containernetworking-plugins-0.8.6-1.fc33 <
containernetworking-plugins-0.8.6-11.1.gitbd58999.fc33
crun-0.14.1-1.fc33 < crun-0.14.1-2.fc33
fabtests-1.11.0-1.fc33 < fabtests-1.11.0rc2-1.fc33
igt-gpu-tools-1.25-1.20200818git4e5f76b.fc33 <
igt-gpu-tools-1.25-2.20200719git9b964d7.fc33
libfabric-1.11.0-1.fc33 < libfabric-1.11.0rc2-1.fc33
slirp4netns-1.1.4-1.fc33 < slirp4netns-1.1.4-4.dev.giteecccdb.fc33
stdair-1.00.10-5.fc33 < stdair-1.00.10-6.fc33
sympa-6.2.56-2.fc33 < sympa-6.2.56-2.fc33.1
kevin
Hrm. It looks like at least some of those issues were transient, yes.
However, two issues are still left:
- libdkimpp-2.0.0-6.fc33 is tagged with f33 and f34 but 2.0.0-3.fc32
is in the f33 repos
- swift-lang-5.2.5-1.fc33 is tagged with f33 and f34 but 5.2.4-3.fc33
is in the f33 repos
Both these packages were built ~two weeks ago, so maybe they were
screwed up by the mass branching somehow?
Fabio