Foundations Team Updates - Thursday 25 May 2023

Project Discussion Foundations
1

The previous status is here: Foundations Team Updates - Thursday 18 May 2023

1 Like
2

Short week due to swap days.

Distro

Netplan

Misc

  • take-home test reviews
  • HR stuff
1 Like
3
1 Like
4

Short week due to public holidays

Rust

1 Like
5

Upcoming boards

  • Install Ubuntu on upcoming boards.

StarFive VisionFive 2

  • Review U-Boot patches

u-boot-nezha

  • Correct debian/README.Debian (LP: #2020692)
  • Add upstream fixes (LP: #2012644)
  • Still needs testing and sponsoring

Greenhouse

  • Review submissions
1 Like
6

Distro

  • armhf time_t work: fixed analysis on 10 packages (although a-c-c runs out of address space on libclang-15-dev), made improvements to the driver script to parallelize work and reduce overhead
  • discussions about which version of openssl to package next (we’ll most likely stick to 3.0 due it being LTS)

Misc

  • Preparing a team application to Ubuntu Bug Control
  • More work towards cleaning the openssl tracker in launchpad
1 Like
7

dotnet

  • Working on .NET preview 4 packaging on Mantic
    • Got a successful offline non-bootstrap build going by adding a few extra prebuilts
    • Working on including these prebuilts on the bootstrap package to re-release it in PPA
1 Like
8

autopkgtest

  • Review of @andersson123’s MP for updating ubuntu.csv on all the machines when opening the archive for a new release.
  • Rebooted the existing autopkgtest-cloud-worker with a zfs compressed /tmp parition (after switching it to use teh ceph volume) and rebalanced the worker services so that we still have the same number of services running but spread across four workers now.
  • Submitted, merged an MP which better clarifies what seed-new-release (part of the arhive opening process for autopkgtest) does.
  • Documented the process for bringing down an autopkgtest-cloud-worker gracefully.
  • Investigation into the hanging network-manager test cases on amd64. It seems to be an issue with the instances not being reachable after running a specific test.
  • Updated the autodep8 code branch on all the test worker units so that a code change is available. Found out the code update wasn’t enough and had to run a make install on the workers.
  • Reviwed @andersson123’s MP delinting the autopkgtest-cloud code.
  • Submitted an MP adding Canonical Ubuntu QA as a point of contact about never run packages.
  • Submitted an MP switching the flavor used in create-nova-image-new-release to autopkgtest.

Ubuntu

  • Add ubuntu-release-upgrader/i386/* to never_run as its deps are unavailable.
  • Submitted an MP documenting the process for updating the ubuntu-release.kpi.u.c site when opening a new release. Merged it after review by @seb128.

Misc

  • Conducted an interview.
  • Spec review of Foundations team bug sync’ing to JIRA.
  • Administrative work
1 Like
9

subiquity / curtin / probert

distro

  • upstreamed fix for kitty (crypto test when RLIMIT_MEMLOCK is too low)
  • last week - merged sysvinit 3.06-4 (thanks @rbasak for reviewing and sponsoring!)
  • last week - fixed python-ansible-pygments when running against pygments 2.15 (thanks @bdrung for sponsoring!)
1 Like
10

Distro

  • usrmerge: Merged usrmerge 35 from Debian (uploaded 35ubuntu1 to mantic)
  • libnvme/nvme-stas: Backport to Ubuntu 22.04 (LP: #2020461)
  • tzdata: Verified SRUs that fix the inconsistencies introduced by the previous SRU (LP: #2017999)
  • dhcpcd5: Sponsored 9.4.1-22ubuntu1 upload to mantic
  • initramfs-tools: Working on replacing dhclient by dhcpcd

+1 week

Remaining part of my +1 shift last week:

  • golang-github-tdewolff-parse: Autopkgtest for golang-github-tdewolff-minify 2.12.4-1 on amd64 fails. golang-github-tdewolff-minify 2.12.4-2 fixed the failure with the new golang-github-tdewolff-parse version. I retried the test with specifying the new version.
  • stress-ng: bug report #2019079 already described the pthread test case failure on armhf and pointed to an upstream commit that fixes it. I tested the fix in a PPA and uploaded 0.15.07-1ubuntu1. Then I noticed that I falsely assuming that the armhf test in my PPA succeeded. The pthread test case did not crash any more, but did not exit any more and it ran into the autopkgtest timeout. I found another stress-pthread upstream commit that fixes a endian issue. Adding this second patch made the autopkgtest succeed. I uploaded 0.15.07-1ubuntu2, reported the success uptream on https://github.com/ColinIanKing/stress-ng/issues/283 and submitted the changes to Debian.

Apport

Other

1 Like
11

Distro

Netplan

  • Added support for dummy interfaces (PR#361)
  • Improvements on the new error handling schema (PR#334)
  • Add support for all types of tunnels in the keyfile parser (PR#360)
  • Netplan 0.106.1 release
  • New autopkgtests for Network Manager and Netplan (MR#443011)
  • Fixed a problem caused by an autopkgtest where an extra default route added by the test was causing connectivity issues in the instance (MR#443566)
1 Like
12
  • dhcpcd5 MIR
    • git-bisect the upstream seccomp fix for ppc64el
    • git-bisect the upstream seccomp fix for new syscall used by glibc
    • Help uploading dhcpcd5 9.4.1-22 to Debian, and write unblock request.
    • Upload dhcpcd5 9.4.1-22ubuntu1 to Ubuntu, sponsored by @bdrung
      The delta includes fixes for Conflict with ubuntu-minimal, and disable autopkgtest for ntpd. These are hopefully will be included when uploading dhcpcd5 10 to Debian after freeze.
  • Merge console-setup.
    I try to split the logic commits with git-ubuntu. But they are many old changes. So I’m still verifying the commits are split correctly. WIP: https://code.launchpad.net/~zhsj/ubuntu/+source/console-setup/+git/console-setup/+merge/443584
1 Like
13

I’m on +1 this week.

Oddly enough, I seemed to have picked time-consuming issues to work on this week, which makes for a short status:

git-annex

Picked it up from last week’s shift. I banged my head against it for a day and a half with a little bit of progress, but no solve in sight, see https://bugs.launchpad.net/ubuntu/+source/ghc/+bug/2019992

matplotlib

This was more productive. Synced from Debian, then dropped because it was holding up the python3.11 transition, then re-synced, the autopkgtests fail in Ubuntu but pass in Debian.

It turns out to have been caused by our pkgbinarymangler PNG optimization pass that changed the color encoding of the icons, triggering a latent bug in the tk backend. There was an upstream fix for it.

See https://bugs.launchpad.net/ubuntu/+source/matplotlib/+bug/2020666

dh-cargo

Retried in bulk the failing tests, both against dh-cargo and migration-reference. Given the nature of the changes, it’s unlikely to have caused regressions in those packages.

r-base

Ongoing. Human error on the Debian side, the new upstream version was uploaded to Debian unstable and then synced. That triggered quite a few autopkgtest regressions due to a change upstream to disallow calling native symbols (???).

Other

  • A couple of technical interviews.
1 Like
14
1 Like
15

Distro

  • Merged openssh from Debian unstable
    • Fixed LP 2017434 “README.Debian.gz instructions for disabling socket activation inaccurate”
  • Working on LP 2020474 “openssh-server-1:9.2p1-2ubuntu1 cannot be installed from active ssh session”
    • This is caused by upgrading from a version of openssh-server that contains the patch from LP 2011458, therefore only mantic is affected.
    • I believe I have fixed the root cause, and I am working on packaging changes to make the upgrade path smooth.
  • Reviewed https://salsa.debian.org/apt-team/apt/-/merge_requests/293 for Julian
  • Merged avahi from Debian unstable (LP 2020383)
  • Working on a batch of armhf header compilation failures
1 Like
16
  • More v3 camera debugging on mantic; no further so far
  • Ubuntu packaging guide meetings & work
  • Working on another batch of armhf time_t packages
  • lvm2 merge (LP: #2020464)
  • Interviews
  • Take home tests
  • Pi meetings
1 Like
17
  • subiquity
    • PR: 1682 - start an endpoint to request the interactive-sections list
  • curtin
    • MP: 443512 - fix issues with handling of extended partitions when running on pre-Focal releases
    • MP: 443511 - start a dedicated compatibilty layer, and use it to detect a util-linux version new enough to enable some features around large sector size usage and sfdisk --no-tell-kernel
    • MP: 443510 - fix issues with writing raw images on pre-Focal releases
    • MP: 443379 - skip some tests for features related to partition resizing pre-Focal with Dos partition table. These could be offered with a newer sfdisk.
    • MP: 443378 - fix an issue where unit tests would pass with some test runners but not pytest, due to some different assumptions about logging
    • MP: 443376 - merge some code cleanup removing python2 things and adding some missing dependencies
  • livefs-editor
    • PR: 37 - fix a dependency, which helps make pipx install work for livefs-editor
  • pc-gadget PR: 95 - fix partition layout
  • greenhouse
1 Like
18

dotnet

ubuntu-packaging-guide

ubuntu-mir

other

1 Like
19

tpm fde:

  • Implemented boot-managed-by-snapd package

UEFI:

  • grub uploads
  • fwupd-efi uploads
  • fwupd/bionic upload

APT:

  • added PPA snapshot bootstrap support to APT 2.7.1
  • released apt 2.6.1 to fix RC bug in Debian
1 Like
20

autopkgtest-cloud

  • Finished linting autopkgtest-cloud repository
  • Wrote script to update ubuntu.csv on workers in cloud environment when adding a new release
  • Finished upstream debian MR to improve logging in autopkgtest (adds timestamps) MR
  • Wrote a script to check for the fastest mirror prior to mojo run to make sure cloud workers use the fastest mirror
  • Working on updating seed-new-release script to retry (connection is flaky and is currently best used in a while true loop to ensure it eventually works without having to manually retry it)
  • Approved and merged a couple of Brian’s MP’s

auto-upgrade-testing-specifications (WIP)

  • kinetic -> lunar ubuntu-unity
  • lunar -> mantic for all other flavours
1 Like