Last week was an interesting week of +1 maintenance - glibc & ruby3.3
transition, clogged builders & queues, firewall misconfiguration for
autopkgtest runners, and so on.
Anyway, the good thing is that the misconfiguration has been fixed and
the builders are not clogged anymore. Hugs to the QA team. :)
Read on for the my week of +1 maintenance -
retrying-autopkgtest-regressions
================================
Throughout the week, with the availability of the runners, I kept
retrying autopkgtest regressions (105+92+71+21 tests triggered).
I also triggered some migration-reference/0 tests for those that still
failed (73+24 tests triggered).
livecd-rootfs/libpff
====================
Whilst I was focused on the devel uploads, ensuring livecd-rootfs/2.765.46
passes all autopkgtests in Jammy was crucial. The s390x tests were failing
because it couldn't open the ~ubuntu-archive/seeds/ubuntu.jammy/STRUCTURE,
apparently. With further debugging, it appeared that libpff was also affected
the same way on s390x for Oracular, amongst other packages. And further
conversations with the colleagues revealed that is was a firewall
misconfiguration for s390x autopkgtest runners in the bos03 region. This
was resolved by the QA team and on re-triggering, things are now passing!
On another note, I also sponsored an upload[0] of livecd-rootfs for Oracular.
ruby3.3 transition
==================
With my Release team hat on, I gave a go-ahead[1] to the Server team to start
the Ruby 3.3 transition. This meant some additional re-triggers of builds
and autopkgtest of various Ruby libraries. Most needed the right set of
triggers and seemed to be passing but a lot of them still need help!
Those who are on +1, please consider helping -
https://ubuntu-archive-team.ubuntu.com/proposed-migration/update_excuses.html#ruby-defaults.
chrony/ruby-sdbm/ruby3.3
========================
We hit an interesting situation with the upload of chrony to Oracular, which
got entangled with the ongoing Ruby transition. One of the Ruby modules
wasn't rebuilt with the right libruby version and therefore the chrony builds
were failing as it was expecting libruby < 3.3. However, after a long
conversation with Andreas and Lucas around this, the rebuilds went through
and chrony built just fine with them. Chrony, in fact, migrated to the
-release pocket, too.
FTBFS of ruby3.3 on ppc64el is a different story altogether. :)
gdisk
=====
LP: #2039594 was reported for gdisk - unexpected behavior with decimal point
in partition size specification. This looked like a bug that upstream fixed
so I was going to forward it to Debian but the maintainer was ready with
the new upstream version upload, which sync'd in Oracular timely and therefore
this bug is fixed. It'd be good to SRU the patch to stable releases though. :)
node-{ssri,yarnkpkg}
====================
This looks like a fallout from the merge we did in Ubuntu - tests seem to
pass in Debian but failing in Ubuntu for all the architectures - looks like
it's trying to access the internet. I've pinged Zixing to take a look and
drive this further. node-ssri can also get a pair of eyes once node-yarnpkg
is unblocked. I've indicated the same to Zixing.
ikwiki
======
Per Graham's mail about the FTBFS in Oracular
(cf: https://people.canonical.com/~ginggs/ftbfs-report/test-rebuild-20241209-oracular-oracular.html),
I looked at ikiwiki as it was
also a problem on the Debian side. The bug report had the root cause and a
patch associated with it. Simon uploaded the fix by the time I completed my
testing so I sync'd the version from unstable to Oracular.
ruby-xmlparser
==============
Per Graham's mail about the FTBFS in Oracular
(cf: https://people.canonical.com/~ginggs/ftbfs-report/test-rebuild-20241209-oracular-oracular.html),
I looked at ruby-xmlparser,
which was FTBFS'ing with GCC-14. The patch was available and simple, so I
uploaded it to Debian and sync'd it to Oracular.
zlib
====
The build of 1:1.3.dfsg+really1.3.1-1ubuntu1 upload was failing on s390x
and some reverse-{build-}dependencies had their autopkgtest failing on
different architectures. I triggered the build and the autopkgtest with
right triggers and they're passing. 3 packages still fail on s390x - osmpbf,
snapd, and zfs-linux. Those are on +1 this week, can you take over, please? :)
simple re-triggers
==================
There were few packages that needed simple rebuilds or autopkgtest re-triggers
(w/ triggers) to get them going -
- python-sphinx-contributors - in -release now.
- ugrep & misspell-fixer - in -release now.
- golang-github-charmbracelet-lipgloss - in -release now.
- blinkpy - in -release now.
- persistent-cache-cpp - in -release now.
- intel-microcode & check-dfsg-status - both in -release now.
- libayatana-common - in -release now.
- calmzip - not blocking ocaml anymore.
- govee-ble - in -release now.
- libtext-csv-xs-perl & restfuldb - in -release now.
- and more which I couldn't keep up with. :)
assisting others
================
I've been assisting the upstream of Gnome Commander to review the bugs in
Ubuntu. Uwe Scholz recently reached out on the mailing list asking for
assistance for the gnome-commander bugs listed at:
https://bugs.launchpad.net/ubuntu/+source/gnome-commander
I've been helping them to update the status of those bugs by pair-triaging
them with Uwe.
needs further debugging (aka handover ;))
=========================================
These are the cases which need further debugging. I didn't have enough time
to investigate -
- python-lark & charliecloud on s390x - still stuck in -proposed. :(
+ needs further debugging - looks like it's querying the internet on
port 443!?
- firewalld - autopkgtest fails on all the architectures. :(
- zlib - read above.
Feature Freeze
==============
We also achieved Feature Freeze[2] for Oracular on 15th August 2024. Whilst
I was on +1, with my Release team hat on, I mailed ubuntu-devel-announce@
about the same. This also means that Debian Import Freeze has been established
and auto-syncs have been turned off.
A gentle reminder to be careful whilst uploading to Oracular now. If your
upload is a feature addition, please file a Feature Freeze Exception (FFE)
bug. The process is documented at [3].
---
[0]: https://code.launchpad.net/~codyshepherd/livecd-rootfs/+git/livecd-rootfs/+merge/471011
[1]: https://lists.ubuntu.com/archives/ubuntu-release/2024-August/006168.html
[2]: https://lists.ubuntu.com/archives/ubuntu-devel-announce/2024-August/001355.html
[3]: https://wiki.ubuntu.com/FreezeExceptionProcess
- u
--
ubuntu-devel mailing list
ubuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
