Foundations Team Updates - Thursday 25 January 2024

Last status can be found at: Foundations Team Updates - Thursday 18 January 2024



  • hiredis ppc64el FTBFS fix was sponsored over the weekend (LP: #2049188, thanks tsimonq2)
    • Both hiredis and kannel have migrated to release
  • Updated python-cssselect MIR template (LP: #2048760)
    • MIR team OK, waiting for security review

shim and GRUB

  • Working on shim 15.8 update for Ubuntu
  • GRUB 2.12-1 is now uploaded and published in Debian Unstable (Thanks juliank for sponsoring).
    • Merge in Ubuntu coming soon


  • Implemented dbx based revocations tests
  • Started doing more validation on arm64
    • I am using the Raspberry Pi 5 as a local arm64 KVM host
    • Fixed multiple arm64 specific bugs
  • Found bug in python-uefivars stopping it from working on AAVMF (edk2 for arm64)
    • Submitted patch to upstream
    • Workaround in ubuntu-boot-test
  • SBAT revocations tests in progress
1 Like


  • Wrote an email and drafted a specification about revoking small signing keys in APT/Launchpad/3rd party repos.
  • Discussed having a field in release files to provide help to users in case of issues, upstream.
  • In merge request apt-team/apt!323, I have added an --audit option and corresponding error log level to apt to flag future deprecations and recommendations for repository owners.
  • Got a report about a new phasing bug in APT that causes Nvidia modules to be kept back for new kernel which ends up with black screens. Have written a fix but not finished the tests yet.


  • Reviewed and sponsored Debian upload grub 2.12-1 for @mkukri


  • Work on livecd-rootfs to enable deb822 sources by default in the image
  • Patch pilot
1 Like


armhf time_t



  • FOSDEM planning and additional paperwork


  • End of my +1 shift. I worked on a couple more FTBFS (see my report on ubuntu-devel)
  • Reviewed some MPs on livecd-rootfs
  • Reviewed a craft-parts related spec


  • Released the 3.2 version in latest/stable
  • Solving LP: #2049695 with PR #178
    • reviewed
    • doing small fixes before merging


  • Some work on designing ubuntu-image rootfs creation mode
1 Like


  • Worked on refining the .NET Snaps specification (FO-151)
    • Included details about the .NET installer tool that will be shipped with the .NET snap to allow for side-by-side version installation
  • Working on a PoC .NET Snap package with the .NET installer tool
  • Attended weekly .NET partners sync meeting
1 Like
  • Doing +1 this week. So far my time has been spent on transitions. No-change rebuilds have resolved most of tinyxml2. ros-ros-comm is still outstanding and entangled in py3.12, I have a patch for that but no proper way to test it. hiredis is the next transition I’m looking at, with no-change rebuilds started. Build failures related to the hiredis no-change rebuild can be seen on my hiredis ppa. Thanks to @ginggs who helped point out an improvement in how I was generating no-change rebuilds (dch -R).
  • python3-launchpadlib (LP: #2050186) - There was a report in IRC with a build log showing a dependency problem. python-launchpadlib uploaded with an additional dependency declared on python3-six. Fix uploaded.
  • curtin - minor fixes then merge Merge into master : cryptoswap : lp:~dbungert/curtin : Git : Code : curtin
  • continued work on mini.iso
  • greenhouse
1 Like




  • hiring
  • sprint sessions
1 Like


  • split our changes into 40 (sic!) patches
  • new Debian version merged via git-ubuntu


  • it is released finally

U-Boot Lichee Pi4A

  • started looking into the MMC driver that is missing upstream

Image for silicon vendor

  • firmware packaged
  • waiting for kernel


  • looking into a bunch of applications
1 Like


  • OpenJDK Security updates

    • openjdk-lts MR sponsored, thanks tsimonq2 (openjdk-17/21 have sync’d)
    • prepared OpenJDK security updates for mantic, jammy, focal and bionic - tested them in a private PPA before uploading to the Security PPA (awaiting results for riscv on focal)
  • Default JDK 21 migration in Debian - submitted fixes to reverse dependencies of openjdk-**

    1. qpid-proton-j-extensions
    2. opencensus-java
    3. zeroc-ice

Distro/Proposed Migration

  • Migration of python3-werkzeug from -proposed (excuse bug):
    • reviews on the lektor/onionshare MPs suggested submitting the same patches with Debian. Created MRs for lektor & onionshare. The onionshare MR has been merged.
    • made improvements to the lektor MP which is still under review


Adoptium Workgroup meeting

1 Like



  • Reviewing take home tests
1 Like
  • Back after several weeks of battling pneumonia \o/
  • Patch piloting on Monday
  • Added mantic and jammy SRUs for LP: #2039873
  • Bumped rpi-imager to 1.8.5 (LP: #2050340)
  • Fixed a build failure in the rpi-imager snap and updated it to 1.8.5 (beta channel)
  • Verified ubuntu-settings backport of clean-up (LP: #2039104)
  • Working on u-boot merge


  • Working on improvements for SR-IOV NICs and better handling of VF LAG activation for Mellanox NICs. (Pending PR)
  • Fixed the generation of the netplan-sriov-rebind.service where we were adding duplicate items in the list of interfaces PR#437
  • Dropped unused code in the SR-IOV logic PR#435
  • Added a couple of missing dependencies to the Coverity workflow PR#436
  • Code review PR#434
  • Investigated why Wireguard tests are failing on ppc64el. It turns out the wireguard module is loading some modules specific for power10 and failing to do so in power9 machines. Seems to be a problem with how the dependencies are managed by the kernel on power. LP#2049924


  • Attending some sprint sessions
1 Like

The biggest part of my week has been a single bug investigation that has lead me down quite a few rabbit holes, including:

  • Pango sources
  • fontconfig sources
  • GObject and how to actually look inside one in GDB (couldn’t figure it out)
  • the Python API of GDB (doesn’t expose symbols per ELF object, sadly…)
  • ltrace (and why it’s actually useless in this case)
  • eBPF (with a bit of time I could write a ltrace-like tool that wouldn’t be limited to calls from the main binary! But that’s too much time)
  • LLDB API → That’s my next hope!

Of course while writing this status report I ran a build against the latest snapshot just to double-check, and naturally the FTBFS vanished.

Computers, what a great invention.




Docs sets

  • WIP: Adding accessibility (pa11y) CI check to docs
1 Like



1 Like


  • refactoring .NET autopkgtests
  • technical discussions about .NET Snaps implementation
    • includes review of FO151 – .NET Snaps
  • preparing presentation for virtual conference talk about .NET on Linux
  • wrote initial draft of FO154 – Alternative package feed for .NET releases not published to the archive
  • attended .NET Security Partners meeting

Ubuntu Packaging Guide

  • attended bi-weekly Sync meeting

Upstream work at canonical/sphinx-docs-starter-pack

  • PR #174 – feat: optional dependency management
    • incorporated technical author feedback
    • got merged :tada:
  • opened PR #179 – feat: remove redirects extension
  • opened PR #180 – feat: remove notfound.extension
  • opened PR #181 – fix: inconsistent help text of Makefile


  • HR Paperwork
1 Like


1 Like


  • Working on making Autoinstall error handling better, primarily looking at the validation side of things this week.
  • Addressing feedback in my proposed spec regarding apport hooks
  • Updates to Subiquity’s Apport hook - MP #454563
  • Added a link to Subiquity’s hosted documentation in - PR #1889


  • Opened a MP to add a new script running-autopkgtests to ubuntu-dev-tools (Originally lp-test-isrunning from lp:ubuntu-server/+git/ubuntu-helpersl)- MP #459002
    • Looking for reviewers on this if anyone has some spare cycles
1 Like


  • Completion and merge of the ignore-list PR and MP, along with a small update to QEMU CLI calling
    Everything working fine, some jobs are back green with a much nicer output for known crashes.


  • Fixed and merged the hardcoded URLs MP
  • Rebased and merged precise code cleanup
  • Did a lot of review, greatly reducing the amount of opened MP
  • Made some progress again in having a working local devel setup (this MP)
  • WIP deploying some bos03-arm64 workers on staging. It’s our first PS6 environment, so we naturally struggle with firewall issues.
1 Like