Last status can be found at: Foundations Team Updates - Thursday 04 January 2024
Distro
- Setup debian-installer building container
- Hit d-i FTBFS, due to usrmerge, https://bugs.debian.org/1060134
- Install & boot custom UEFI d-i image
- component-mismatch: lxml vs python-cssselect investigation, https://pad.lv/2048760
- investigate removal of dmraid (seeded & Ubiquity recommends)
- migration: driving python3-werkzeug forward
- Bug #2048769 “Adopt to new python-werkzeug 3.0 API” : Bugs : python-werkzeug package : Ubuntu
- unlocking pytest-httpserver, flask-login, flask-dance
- patch piloting, Patch Pilot Hand-off 24.04 - #45 by slyon
- systemd-hwe, netdata, rsyslog, ignition-cmake, cryptsetup, gnutls28
- sync python-ulmo after our delta was accepted (cc @ogayot)
Netplan
- review OVS autpkgtest fix, https://github.com/canonical/netplan/pull/431
- review C unittest quirks, https://github.com/canonical/netplan/pull/432
- merge Docs Starter Pack, https://github.com/canonical/netplan/pull/429
- merge Docs Howto cleanup PR, https://github.com/canonical/netplan/pull/430
- review, merge & sponsor netplan.io 0.107.1 (unstable & noble)
- prepare API & ABI drop/cleanup PR, https://github.com/canonical/netplan/pull/400
Misc
- hiring
2 Likes
OpenJDK
TCK-17
- Added support to run individual test targets. There are ~150 significant targets.
- Worked on setting up a VNC server for tests needing DISPLAY. Reduced failure count by 20.
- A few more minor fixes to the scripts brought down the total failure count to 2 (on amd64 and arm64)
Debian openjdk-*
- Fixed a bunch of minor issues in the copyright generator for openjdk (MR # 88)
Java FIPS provider prototype
- Started adding a Java layer to the openssl wrapper written last year
- Added a Java API and tests for DRBGs.
CRaC
Misc
- Adoptium workgroup meeting
- Meeting with Security team to understand their openjdk-11-fips review strategy
1 Like
out sick a day this week
- sbuild vs dpkg - sbuild autopkgtest correctly discovered a regression in dpkg in terms of the compression level used for zstd (LP: #2048137), which appeared as diffoscope reporting that a file matched what was expected in terms of uncompressed contents but that the compressed bytes were different
- sbuild autopkgtest also a bit broken at the moment due to side effects of restrictions on user namespaces (LP: #2048788) - while I spent some time on this one I don’t have a great answer for what to do about the sbuild ADT, especially as it was useful in finding the above bug
- refamiliarizing myself with cryptoswap work from before break, expect to be opening a curtin MP on this today
1 Like
Distro
- Followed up on some +1 carryovers
- Everything for Ubuntu was sponsored.
- Some Debian NMU diffs still sitting around.
- Fixed ignition-cmake x cmake FTBFS (LP: #2048518)
Boot
-
Working on adding the ability to test shims before ms signing
- Tested modification for shim-signed to build without MS signed shim in a PPA
- Implemented signature detection in ubuntu-boot-test to figure out if it’s encountering a production key vs PPA key scenario on the fly
- TODO: update specification with this information
-
Initial merge of GRUB 2.12 final for Debian: https://salsa.debian.org/mkukri/grub/
-
Miscellaneous bootloader work and testing
1 Like
Distro
- lomiri-thumbnailer: Synced from Debian after the Ubuntu delta was merged in Debian (see Debian bug #1031508)
Apport
- bug triaging session with @schopin
- feat: support reading report from systemd-coredump
- ci: also measure branch coverage
- test: improve debugging failed test_recent_syslog_overflow
- Reviewed add --minimum-java-release option to the build command
- test: add test_crash_suid_dumpable_debug
- apport: fix UID in report filename for suid programs
- test: make branch coverage report stable
- refactor(ui): simplify symptoms handling in run_symptoms
- Draft: feat(apport): add --from-systemd-coredump
Sponsoring
- Sponsored merge of usrmerge 38ubuntu1 for @ravi-sharma
Patch pilot
proposed migration
- Uploaded pycryptodome 3.20.0+dfsg-1 to Debian unstable to fix Python 3.12 support and autopkgtest on non-x86 architectures
1 Like
Subiquity
- PR reviews
- spec work
- aiohttp security discussions
Distro
- retriggered autopkgtests for libcgi-application-plugin-authentication-perl to unblock libcgi-pm-perl
1 Like
socat
- update to new Debian release (LP #2048408)
tang
- fix timeouts in tests (LP #2048479).
U-boot
- investigate options for supporting Lichee Pi4A
1 Like
installer
- merged probert PR to get NVMe controllers listed in the probert output
- opened curtin MP adding partial support for NVMe-o-TCP drives
- subiquity PR to be opened
- spec work
distro
- dbus
- 1.14.10-3ubuntu1 : dbus package : Ubuntu now in -proposed. Thanks @tsimonq2 for sponsoring!
- currently blocked by apport which has failing autokpgtests
- staged requested changes for next dbus merge in a branch in launchpad
- started investigating
netplan.io
blocking wpa migration because of autopkgtest failures. Retries don’t see to help. I filed bug 2048388. I also added some debug information tonetplan.io
testsuite which I uploaded to a PPA.
1 Like
Go
- Update Go 1.21.6 and 1.20.13 releases. FIPS packages and snap will be updated next day.
- Investigating the Go 1.22 rebuild results. Will try to fix the FTBFS packages next week.
1 Like
.NET
- preparing 8.0.101/8.0.1 micro-release for jammy
- initial review of FO151 - .NET Snaps
- triaged LP: #2048926 - Not getting listed in dotnet --list-runtimes so facing an error missing 6.0 runtime
- attended .NET Security Partners meeting
Ubuntu Packaging Guide
-
writing articles
-
reopened Discussion #30 - Should we add the sphinx-tabs extension?
-
reviewed & merged PR #48 - Misc. minor fixes and updates
-
canonical/sphinx-docs-starter-pack
- Issue #164 Inconsistent help text of Makefile
- working on optional dependency managmenet
1 Like
Distro
- Uploaded systemd-hwe for Jammy, Noble, and Mantic (LP: #2046687 and LP: #2045621)
- SRU verification for systemd SRUs
- Sent a patch to fix a test case in systemd: https://github.com/systemd/systemd/pull/30854
- Generally investigating systemd test failures in LXC
- Worked on fixing issues caused by recent AppArmor changes: Merge into ubuntu/devel : ubuntu/devel : lp:~enr0n/ubuntu/+source/apparmor : Git : Code : apparmor package : Ubuntu
- Uploaded ubuntu-release-upgrader: 1:24.04.4 : ubuntu-release-upgrader package : Ubuntu
- Had conversations with Ubuntu Frame folks about systemd service ordering etc.
- Sponsored rsyslog for @xypron: 8.2312.0-2ubuntu1 : rsyslog package : Ubuntu
Misc
- Reviewed several take home tests
1 Like
This is my first work week coming from the end-of-year holiday break.
dotnet
- Worked on writing the .NET Snaps specification (FO-151)
- Working on building PoC for the .NET Snaps per the specification
- Attended weekly .NET partners sync meeting
1 Like
autopkgtest
- Further worked with IS to resolve a networking issue causing amd64 tests not to run see RT for details regarding the issue.
- Submitted an MP adding support for a package configuration file which specifies additional architectures for packages to run tests on for ESM releases.* Tested the above changes using ubuntu-advantage-tools and armhf for xenial.
- Cleaned up after some tests which had failed due to a lack of free space on the autopkgtest-cloud-worker unit.
- Readding / rebooting lxd remotes which had gone AWOL.
- Helped @andersson123 debug an issue with PPA tests running in production. Reviewed and merged the resulting MP.
- Submitted, merged, cowboy’ed an MP adding s390x to the list of allowed architectures for which tests run on Xenial.
- Pinged IS regarding restarting neutron for s0lp4. RT 155441 is the ticket about getting this fixed for reals.
- Updated the branch of autopkgtest used on the autopkgtest workers to branch with version number 5.20.
- Testing @hyask’s development version of prod-proposed-migration.
- Updated my MP which hides warnings from apt to use @vorlon’s “horrible shell pipeline” which better manages stderr.
Error Tracker
- Requested a staging environment for the Error Tracker in PS5 since it hasn’t magically happened post Riga. It was rejected and I discovered there actually is one but on the foundations-bastion.
ISO Tracker
- Debugged the sync-lp-bugs script on the tracker which involved playing with the postgresql configuration and authentication.
- Submitted, merged an MP which ports sync-lp-bugs to python3 and modifies the configuration file for the script.
- Added cronjobs so that the sync-lp-bugs script runs regularly.
Ubuntu
- Encountered a crash with listadmin again, found a patch upstream, updated the debian bug, created an Ubuntu bug and uploaded the fix for noble.
Misc
- SRU team member rotation.
1 Like
Rust
- Published Rustup Snap; alias request is still pending
- Prepared versioned Rust 1.75 package for transitioning to the versioned package scheme
- Preparing
cmake
backport for Rust 1.74+ backport (due to LLVM 17 requirements)
1 Like
Distro
- coming up with a list of packages (WIP) that are servers and have custom crypto configurations (mostly about TLS right now)
- also creating a testsuite for that
- force-disabled DTLS 0.9 and 1.0 in gnutls system-wide ( Merge into ubuntu/devel : noble-gnutls-disable-dtls-0.9-and-1.0 : lp:~adrien-n/ubuntu/+source/gnutls28 : Git : Code : gnutls28 package : Ubuntu )
- re-did my openssl SRU with comments taken into account and worked with Simon as he was reviewing it ( Bug #2033422 “openssl: backport to jammy “clear method store / q...” : Bugs : openssl package : Ubuntu )
- specification work for crypto config
- working on disabled TLS 1.0 and 1.1 in rabbitmq
Armhf time_t
- more packages
- new analysis run, storing results now and then I need to diff and share that again
Misc
- attending FOSDEM and did all the preparations for that
Proposed migration
- forgot about python-secretstorage vs vorta but AFAIU it was fixed in Debian
1 Like
- Released apt 2.7.8 to fix valgrind crash (not running it on armhf anymore) and merged open stuff
- Released apt 2.7.9 to fix regressions from merges only visible on infrastructure (fs ordering dependant)
- Weird stuff
- Meetings
- Patch pilot tomorrow, probably should move that again
1 Like
Distro
- Investigated llvm-toolchain-* autopkgtest failures. There is an interesting correlation with the new kernel (6.6 and 6.7) on Noble. Binaries built with address sanitizer are crashing (also with GCC). LP#2048768
Netplan
- Finished the 0.107.1 packaging and @slyon uploaded it.
- Reached out to the UX team for
netplan status --diff
testing - Wrote a small topic about security for the Netplan docs PR#433
- Address the issue with autopkgtests and the new systemd 255 PR#431
- Proposed a better way to run the C unit tests without having to include other C files in the test files PR#432
- Added support for generating random MAC address also for networkd PR#427
- Investing autopkgtests failures with the new version of Netplan
1 Like
Documentation
Packaging Guide
- PR #48: Misc. fixes and updates (templates, theme, …)
Installation Guide
- PR #1883: Misc. docs updates and fixes (build warnings, mark-up)
1 Like
Autopkgtest-cloud
-
Fixed autopkgtest-cloud devel environment stuck in
Waiting on RabbitMQ to configure vhost
: issue in service-bundle withcookies
- Investigate the diff between
autopkgtest-cloud
anddebci
: they’re just two very different codebase, not even in the same language: very hard to share anything here. - WIP fixing the “unshare” test for
autopkgtest
package: dependency ondebian-archive-keyring
missing. Upload to PPA and test in prod env, thennoble
seems to be pretty broken, testing again onmantic
(still WIP as of now). - WIP writing tests for
autopkgtest
's handling of lack of free space, related to this MP and this bug. This is a race condition that is pretty hard to reproduce reliably. My current test hangs only 1 time out of 10 or 20 tries, so too low for working with that easily.
Auto upgrade testing
-
MP to activate
RELEASE_UPGRADE_NO_FORCE_OVERWRITE
to test upgrades withoutdpkg --force-overwrite
and catch more bugs. Nothing new appeared after that.
curtin
- Fixed infrastructure (proxy) issue with the Jenkins job
curtin-vmtest-daily-f
. Now it’s able to run, and last time, two tests ran into timeout errors, thus will need follow up work.
1 Like