The criu test failure seems to be caused by a bug in mumble's scriptlets:

>>> Running %posttrans scriptlet: mumble-server-0:1.4.287-11.fc45.x86_64
>>> Non-critical error in %posttrans scriptlet: mumble-server-0:1.4.287-11.fc45.x86_64
>>> Scriptlet output:
>>> rmdir: failed to remove '/etc/murmur': No such file or directory

I'm not sure why rpmdeplint.functional is missing for perl-Alien-ProtoBuf.

rmdepcheck notes one missing dependent rebuild:

Dependencies of other packages that would be BROKEN by the tested packages:
package: mysql-connector-java-1:8.0.30-21.fc44.noarch from https://kojipkgs.fedoraproject.org/repos/f45-build/latest/x86_64
  mvn(com.google.protobuf:protobuf-java)

Oh, rpmdeplint was skipped because there are more than 20 builds in the update.

I do wonder if we could stop doing that, now I made it faster.

Looking into mysql-connector-java, the thing it requires was provided by protobuf-java , which you seem to have simply removed in this update because "Can't get it to build". That seems questionable.

criu test now appears to have some issues allocating memory:

========================= Run zdtm/static/maps09 in h ==========================
Start test
./maps09 --pidfile=maps09.pid --outfile=maps09.out
make: *** [Makefile:523: maps09.pid] Error 1
 Test zdtm/static/maps09 FAIL at ['make', '--no-print-directory', '-C', 'zdtm/static', 'maps09.pid'] 
Test output: ================================
19:25:41.868:    70: ERR: maps09.c:20: Failed to mmap 4 Mb anonymous shared memory (errno = 12 (Cannot allocate memory))
19:25:41.868:    69: ERR: test.c:320: Test exited unexpectedly with code 1

 <<< ================================
========================= Run zdtm/static/maps10 in ns =========================
Start test
./maps10 --pidfile=maps10.pid --outfile=maps10.out
The test failed (0, 1)
make: *** [Makefile:523: maps10.pid] Error 1
 Test zdtm/static/maps10 FAIL at ['make', '--no-print-directory', '-C', 'zdtm/static', 'maps10.pid'] 
Test output: ================================
19:25:42.402:     4: ERR: maps10.c:26: Map failed (errno = 12 (Cannot allocate memory))
19:25:42.402:     3: ERR: test.c:320: Test exited unexpectedly with code 1

@adrian , can you see what's going on here by any chance? It seems the tests passed on your most recent standalone update a month ago...

We plan to start the Python 3.15 rebuild tonight. I would feel much more comfortable if this landed before that.

I guess we could try and see if the criu tests fail outside of this update, somehow. But that still leaves the missing java interface.

I added protobuf with java subpackage and rebuilt mysql-connector-java with the new protobuf-java

The tests for mysql-java-connector are failing with expected ExclusiveArch "aarch64 ppc64le s390x x86_64 riscv64" but was "" I have no idea why it is there. What about the criu? If the test is since previous update I am inclining to waive it out.

I have just discussed the failing "mysql-java-connector" test with @mkoncek and it is ok to waive it. The tests are made for single package update and will not wokr for side-tag builds with more packages.

I'm trying to see if I can reproduce the criu failure with older protobuf. It's a PITA to get it running.

Meh, it's rather difficult to reproduce the test process with any confidence, but I think it's probably okay. What I wound up doing is setting up a Rawhide VM and running the python3 ./zdtm.py run --criu-bin /usr/sbin/criu -a --ignore-taint --keep-going command from the test suite (in a checkout of the package, after building criu within the source dir...), then update with all the packages from this update and doing the same thing.

On the 'before' run, the tests file_locks01(uns) and rseq01(h) failed; on the 'after' run, those tests failed and also socket-tcpbuf-local failed. But I think that's likely a flake; I think it also failed on an earlier 'before' run (when I wasn't properly logging the output). Neither run shows mmap failures.

So...I think it's probably OK to waive the criu failure? But it'd still be great if @adrian could look at it.

fwiw, libphonenumber-8.13.55-10 silently changed abi. e-d-s->evolution ails to start

evolution: symbol lookup error: /lib64/libebook-contacts-1.2.so.5: undefined symbol: _ZN4i18n12phonenumbers11PhoneNumberC1EPN6google8protobuf5ArenaEb