The previous status is here: Foundations Team Updates - Thursday 25 May 2023
Archive
- NBS processing of pep8 which has been replaced by pycodestyle
- component-mismatches management
Release
- Misc assistance with bionic move to ESM
- Fixed issue identified with the pending->current promotion of daily images not working after the server move
SRU
- Full SRU rotation on Friday
Distro
- Ongoing work on the time_t migration
- Met with Robie to discuss how to best approach contributing new features to git-ubuntu
- have started working on a prototype for
git-ubuntu build
to wrapdpkg-buildpackage
- used it in anger a couple of times
- have started working on a prototype for
- Drafted an SRU exception for
ubuntu-dev-tools
, found out to my chagrin there already was one. But the new one is a bit lighter weight - SRUed
debootstrap
andubuntu-dev-tools
everywhere to address https://bugs.launchpad.net/bugs/2020530 - Sponsored SRUs of
util-linux
andopenssh
- Netplan reviews
Misc
- Candidate interview
- Reviewing a job description
- curtin
- LP: #2003745 - finalizing fixes for older series support
- subiqutiy
- investigate issues in acquiring autoinstall data from cloud-init
- PR: 1682 - add endpoint for getting the list of interactive sections
- LP: #2019856 - sponsor util-linux lscpu improvement for @mfo
- greenhouse
- specification, github, and launchpad reviews
Short week, was off Mon/Tue
apt:
- APT 2.7.1 with PPA snapshot support
boot:
- grub2-unsigned 2.06-2ubuntu17 and 2.06-2ubuntu14.2 SRUs
- fwupd-efi 1.4 SRUs
- boot-managed-by-snapd/mantic is NEW
misc
- Greenhouse
- armhf time_t work
Partial week due to public holidays.
glibc:
- Prepared SRUs for Jammy and Focal, including a bit of bug grooming.
- In the middle of rebasing 3 years worth of Ubuntu-specific changes on top of Debianâs packaging. Fun times!
+1 maintenance
I looked at packages that are FTBFS and stuck in proposed migration due to missing build.
-
gitlab-ci-multi-runner
The package has long history of ftbfs both in Ubuntu and Debian.
Request removal https://launchpad.net/bugs/2021461
-
delve
bpftool needs to match running kernel version. Not possible to have a
working bpftool command in Ubuntu build environment. (Disclaim: I made the requirement for bpftool in Debian. it works in Debian because the bpftool is packaged differently. I would revert that change in Debian if no solution found when Debian is not in freeze.) -
librnd + camv-rnd + pcb-rnd
camv-rnd and pcb-rnd are in dep-wait status, it needs librnd > 4.
Itâs a small library transition. So I just ask @ginggs to kick off the transition.Also ask for revoking the blacklist for sch-rnd since it has stable release now.
https://launchpad.net/bugs/2007172 -
perl related
Several packages FTBFS with same test failure:
event webhook_request of session KGB for DIR handler â^/webhook/â. As
reported by Kernel: âNo such file or directoryâ, perhaps the session name is
spelled incorrectly for this handler? at /usr/share/perl5/POE/Session.pm
line 483.Iâm still debugging that.
-
godot
FTBFS in Debian too https://bugs.debian.org/1031132
Imported the bug to launchpad and add a update-excuse bug for reference. -
libs3
FTBFS on ppc64el only. Caused by -Werror=stringop-overread. Looks like the
difference is -O3 vs -O2 build flags between Ubuntu and Debian.Patch attached at https://launchpad.net/bugs/2021564
-
libtgowt
FTBFS on riscv64, but previous version not. The new version needs upstream
explicit support in its build config. However this library is only for
building telegram-desktop.Request removal https://launchpad.net/bugs/2021567
-
kickpass
FTBFS on amd64 due to LTO. Caused by a pie patch which is added by the Debian
maintainer (Can be safely dropped).Patch attached at https://launchpad.net/bugs/2021577 and forwarded to
https://salsa.debian.org/debian/kickpass/-/merge_requests/1 -
pushpin
FTBFS on ppc64el, riscv64, s390x. Same on Debian, and Debian has removed the packages on these architectures.
Request removal https://launchpad.net/bugs/2021594
Distro
- initramfs-tools:
- Port the net autopkgtest to the common test framework. This drops depending on downloading a cloud image from the Internet and reduces the execution time from 3:19 min down to 0:57 min: https://salsa.debian.org/bdrung/initramfs-tools/-/commits/debian%2F0.142ubuntu3?ref_type=tags
I am waiting for @waveform to come back to review these changes before the upload. - test: Fix too small ext2 block count
- Several improvements for debian/tests/test-common
- Working on replacing dhclient by dhcpcd
- Port the net autopkgtest to the common test framework. This drops depending on downloading a cloud image from the Internet and reduces the execution time from 3:19 min down to 0:57 min: https://salsa.debian.org/bdrung/initramfs-tools/-/commits/debian%2F0.142ubuntu3?ref_type=tags
armhf time_t
- Resolved armhf_time-t compilation failures: https://salsa.debian.org/vorlon/armhf-time_t/-/merge_requests/44
- check-armhf-time_t: Ensure that eatmydata, killall, and ctags are installed
- zziplib: zzip/fseeko.h: Fix closing extern C bracket
Distro
- Worked on util-linux focal SRU: LP 2015355 âPlease backport patches for false atari partition detection to Ubuntu 20.04â
- Worked on LP 2020474 âopenssh-server-1:9.2p1-2ubuntu1 cannot be installed from active ssh sessionâ
- The issue seems to be resolved now, but still waiting on migration
- Also made a follow-on fix to LP 2011458 so that we avoid this bug in Kinetic and Lunar
- armhf time_t batch
short week (3 days off)
subiquity
- spec work and planning for ubuntu-drivers integration (OEM related)
- patch for ubuntu-drivers --no-oem merged (uploaded in mantic). SRU to be initiated
- merged mypy script to compare two branches
distro
- had a first look at dbus merge
dotnet:
- still working on release automation tooling
unzip:
- LP: #2021565 Merge 6.0-28 into mantic (reviewed by Steve Langasek (vorlon))
- LP: #1004368 New -> Inclomplete (could not replicate in recent versions)
- LP: #1255640 Confirmed -> Incomplete (could not replicate in recent versions)
- LP: #1087977 New -> Incomplete (not enough information)
- LP: #1099587 New -> Incomplete (could not replicate in recent versions)
- LP: #1230233 New -> Incomplete (could not replicate in recent versions)
- LP: #1405676 New -> Incomplete (could not replicate in recent versions)
- LP: #1434789 New -> Incomplete (could not replicate in recent versions)
- LP: #393987 (Re-)Confirmed (for recent Ubuntu Versions)
-
LP: #2022000
- reported for jammy
- later marked as duplicate of LP: #1429939
- LP: #1429939 (Re-)Confirmed (for recent Ubuntu Versions)
- LP: #1947451 marked as duplicate of LP: #13830
- LP: #1691636 Confirmed -> Invalid (intended behaviour)
-
LP: #1898445
- New -> Confirmed
- affects: unzip -> zip
- LP: #1898407 New -> Invalid (bug was already fixed at time of reporting)
- LP: #1998073 New -> Invalid (could not be replicated)
- LP: #1958437 New -> Invalid (not a bug)
- LP: #1519905 New -> Incomplete (version no longer deployed / could not replicate in recent versions)
- LP: #1361740 Confirmed -> Incomplete (could not replicate in recent versions)
- LP: #192330 importance: Undecided -> Wishlist
- LP: #220654 (Re-)Confirmed (for recent Ubuntu Versions)
- LP: #1824530 Confirmed -> Fix Released
Distro
- Reflected on how to beging implementing systemd-wide crypto configuration without breaking everything (lots of time spent evaluating TLS test clients); PoC to begin this month I think
- Looked at the cdebconf merge but mostly spent time better understanding it better
- Launchpad stuff wrt openssl tickets
armhf time_t
- Fixed batch 11 except for libclang-15-dev: https://salsa.debian.org/vorlon/armhf-time_t/-/merge_requests/38
- because we canât analyze all of its headers at once (too big); WIP: start introducing virtual packages to split the work: https://salsa.debian.org/vorlon/armhf-time_t/-/merge_requests/45
- WIP parallelization of a-c-c jobs for a given package (expect 60% speedup on large packages): https://salsa.debian.org/vorlon/armhf-time_t/-/merge_requests/37
Misc
- Prepared Foundations team application to Bug Control
Distro
- Worked on a batch of time_t related fixes for armhf
Netplan
- Code review for a contributor from the community related to wpa_supplicant configuration. We are using the wrong configuration key when selecting the frequency for adhoc networks PR#363
- Improvements on coverage for memory leak tests PR#365
- Merged the error handling improvements PR PR#334
- Investigating a problem with Wifi adhoc configuration on the Raspberry PI where it will not work out of the box with netplan.
- Continued the work on the Netplan Diff spec
Network Manager
- Investigated and fixed a couple of issues related to the netplan integration MR#443876
autopkgtest
- Resolved an issue with updating the autopkgtest-cloud-worker charm in staging by updating the charm to pin the version of pyaml and submitted an MP with that change. This also involved wrestling with juju a wee bit.
- With @paride resolved another issue updating the autopkgtest-cloud-worker charm in staging this time by pinning PyYAML. See the MP for some more details.
- Updated the autopkgtest-cloud-worker charm in staging and production to revision 156 which includes fixes for a bug regarding images potentially having packages from -proposed in them and a bug regarding a trailing dot in FQDN of hosts in PS5.
- Submitted, merged an MP which documents the process for updating the autopkgtest and autodep8 code in the autopkgtest environment.
- Submitted an MP updating the mojo/service-bundle by removing the autopkgtest-cloud-worker-lrg units and moving test runners to the autopkgtest-cloud-worker units.
- Removed autopkgtest-cloud-worker-lrg units from production as they had been replaced by autopkgtest-cloud-worker units and were no longer being used.
- Pinged IS regarding restarting neutron on s0lp4 in scalingstack-bo02 as the s390x queues were slowly processing again. See RT 155441 for details.
- Submitted a request to have a staging proposed-migration setup in PS5 on the ubuntu-qa bastion. Also reported a bug about the form.
Ubuntu
- Retried autopkgtests which had failed for stable releases using
retry-autopkgtest-regressions
.
Misc
- Scheduling / coordinating team holidays.
- Reviewing candidate applications.
- Reviewed and closed some outstanding RTs that were no longer valid.
dotnet
- bootstrap package for amd64 ready in PPA
- working on arm64 bootstrap to add as dependency to final package
distro
- working on adduser merge
Rust
- Finishing up the
cargo
andrustc
source package merge- The full rebuild test concluded, and no regression was detected
- You can find the built result in https://launchpad.net/~liushuyu-011/+archive/ubuntu/rust-updates-1.68-unify
Distro
- currently investigating the network-based tests for the
fwupd
proposed migration
ca-certificates-java:
- tested and added patches for lts versions of LP: #2003750
openjdk-21:
- investigating jtreg failures of test build
- jtreg 7 was sponsored by @vorlon and now we can make a release of openjdk-21 rampdown phase 1 (when it gets tagged next week)
openjdk-lts:
- prepared a build to test LP: #2021566. Unfortunately without proper reproduction steps it is more of a guess.
chiselled-jre 17:
- doing performance and size benchmarks of a new chiselled Java 17 container (from this branch)