On 08/21/2014 09:17 AM, Dan Horák wrote:
On Wed, 20 Aug 2014 17:27:39 -0600
Kevin Fenzi <kevin@xxxxxxxxx> wrote:
Just thought I would drop a quick note to the list about the status of
various things, at least as I know them. ;)
First, astute readers will note that we were scheduled to go into
freeze for f21 yesterday and that we didn't do so. This was discussed
at the FESCo meeting today and we decided to go into Alpha freeze the
day after we have a viable Test Compose for f21. Going into freeze
before we have that would make it harder to get there, and just
slipping a week for freeze might not be needed.
The recent mass rebuild for glibc/gcc fixes finished monday. However,
it was discovered that there was a rpm bug present for the last ~1300
or so builds in the mass rebuild. This bug would cause packages that
override provides/requires to just not add their requires at all.
https://bugzilla.redhat.com/show_bug.cgi?id=1131892
I have just finished scraping the logs for those ~1300 f21 builds and
identified the packages affected. I will be rebuilding them in f21 and
rawhide. Happily it isn't too many packages. Packages outside of the
mass rebuild may also be affected, maintainers are advised to
doublecheck their build.log if they override provides/requires and
built anything the last few days.
Was about to post this separately but since there's already a relevant
thread...
There's another package generation affecting bug, present in
4.12.0-0.beta1.3 and 4.12.0-0.beta1.4. Any packages containing triggers
which have been built with those versions, need to be rebuilt with
rpm >= 4.12.0-0.beta1.5. See
https://bugzilla.redhat.com/show_bug.cgi?id=1131960 for details and
recipe for detecting affected packages.
the fix is in rpm-4.12.0-0.beta1.4.fc22 if I see correctly, but what
are the broken rpm versions? I would like to skip them on secondary
arches. We don't have replicate also all the bugs :-)
Avoid all rpm 4.12 beta1 versions prior to 4.12.0-0.beta1.5.
Oh and apologies for the trouble caused to everybody :-/ That there were
such package generation affecting regressions is unfortunate to begin
with (caused by insufficient test-suite coverage) but to coincidence
with not just one but two mass-rebuilds...
- Panu, with brown paperbag firmly applied -
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/devel
Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct
