Last week’s update is available at https://discourse.ubuntu.com/t/foundation-team-updates-thursday-09-june-2022
- Added bzr support to the germinate state
- Change Seed URLs to a list to allow germinating from multiple URLs
- Increased test coverage to 100% for germinate state and write tests for multiple functionalities
- Add list of snaps from germinate output to eventually pre-seed
- Discussed some other image definition changes per @mwhudson PR comments
- Opened pull request for germinate state
- Start work on bootstrap state to call
debootstrapand mount filesystems
- Removed even more references to livecd-rootfs and updated the github CI to run on jammy
- proposed-migration: pkg-js-tools vs libjson-perl
- Reviewing gsasl MIR TODO tasks
- Opened pull request upstream in invopop/jsonschema to address a bug seen in ubuntu-image testing
- Uploaded 2.06-3 to unstable with the security fixes. This finally unblocks migrating 2.06 to testing, and we get to work on merging stuff back to Debian once this has migrated
- Some discussions about maybe having 2.06.1 releases
- Sought clarification regarding the behavior of SBAT policy updates
- Analysed phased updates related regression in LP: #1978125, sketched out some approaches on IRC
- reviewed an internal spec
- reviewed https://github.com/canonical/ubuntu-advantage-client/pull/2118
- systemd SRU verification + migration
- systemd-oomd discussions on email@example.com, https://pad.lv/1972159
- MIR review: lerc, https://pad.lv/1977551
- netplan ‘status’ spec, version 2, pending review
- LP bug grooming, cleared all
nplanbugs on LP \o/
- finalized IP6-GRE-TAP tunnel key PR#274, https://pad.lv/1966476, pending review
- netplan.io v0.104 merge in Debian, pending review
- working on InfiniBand (IPoIB) support, PR#283
- short week (only 3 days)
- some Canonistack investigation wrt. cRT#150658 & cRT#131289
- some partner product discussions with the Desktop team
- PR: 1300 - provide a whole new implementation to fix this problem (LP: #1975629) with environment variables and the debug shell. Environment variables modifed for running subiquity in the snap first have their unmodified values set aside, which we restore for the shell.
- PR: 1308 - fix internal usage of free space during partition creation. A prerequisite fix for creating multiple partitions in a single Gap of available space.
- PR: 1309 - several utils for working with Gap objects.
- PR: 1312 - fix issue with loading an autoinstall file directly from the ISO
- PR: 1313 - add utility to find the largest disk gap across several disks
- PR: 1316 - gather up a few size related improvements - create partition is now handed a gap of appropriate size to partition into, and adjust the boot and uefi partition scaling functions to be able to operate when using part of a disk.
- PR: 1318 - remove an assignment to underscore that has the potential in the future to conflict with gettext.
- (no PR yet) - continue work on an autoinstall directive that allows for installing into an existing gap when doing guided partitioning, similar to d-i ‘biggest_free’. Works enough to handle a first VM test case. Most of the above PRs are part of this.
- Specification work
- Review: PR: 1310, 1311, 1314, 1315, 1317, 1319
- Short week, Thursday and Friday off
- FTBFS reports
- Proposed migration (there weren’t enough Foundations packages to go around last week, so looked at a couple of stuck universe packages) :
- Investigation into Bazaar to GIt migration of repos used in the infrastructure and packages in main
- Merged cython (TIL universe package)
- Archive-admin training
- Spec for Ubuntu Pro (attach using Ubuntu One) sent out for review.
- Focused on implementing most screens from the current Ubuntu Pro mockups. Most of the logic is yet to be implemented though.
- Merged PR 1307 - validation of subfields affect parent form.
- Created PR 1314 - ability to customize the label on buttons for error message boxes (merged)
- Created PR 1315 - to fix how visual overlays in Ubuntu Pro screens are closed. Changes were requested -> TODO.
- Investigated LP: #1978139 - Subiquity crashes after refresh for clients connected through SSH
- merge of keyutils uploaded (thanks @bdrung for sponsoring!)
Uploaded apport 2.21.0-0ubuntu1 to kinetic, which had two failing test cases. I reproduced them, fixed them, and uploaded 2.21.0-0ubuntu2.
- sponsored libcloud 3.4.1-5 (LP: #1977320) for @enr0n
- reviewed https://salsa.debian.org/hpc-team/opensm/-/merge_requests/2 -> discussion result: do not merge this change, but figure out why the logs were spammed
- Move onetbb from long_tests to big_packages: https://code.launchpad.net/~bdrung/autopkgtest-cloud/+git/autopkgtest-package-configs/+merge/424741
- Reviewed sbuild merge (LP: #1974177), forwarded some changes upstream:
- Run unshare autopkgtest against target release: https://salsa.debian.org/debian/sbuild/-/merge_requests/19
- Set NO_PKG_MANGLE=1 when building dummy packages: https://salsa.debian.org/debian/sbuild/-/merge_requests/20
- Revert “debian/tests/build-procenv: add deb-src repository”: https://salsa.debian.org/debian/sbuild/-/merge_requests/21
- Reproduced LP: #1976257 and forwarded to Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1012856
- uploaded ubuntu-dev-tools 0.190 to unstable
- aptdaemon autopkgtest fails with iso-codes/4.10.0-1 -> cannot reproduce locally in QEMU runner, triggered retest
- proposed migrations:
- collectd: 5.12.0-9 builds -> closed blocking bug #1960612
- kcollectd: 0.12.0-1 builds -> closed blocked bug #1960612
- xdg-utils-cxx (233 days old): FTBFS due to missing symbols -> uploaded xdg-utils-cxx 1.0.1-3ubuntu1 and 1.0.1-3ubuntu2, forwarded patch to Debian bug #1004109
- golang-github-komkom-toml: armhf failure -> 32-bit issue -> Debian bug #1004048 -> forwarded upstream: https://github.com/komkom/toml/issues/5
- Reviewed, commented, and closed https://bugs.launchpad.net/ubuntu/+source/libsigrok/+bug/1958713
- synced golang-github-prometheus-common 0.32.1-8
- lintian-brush 0.126 FTBFS -> debmutate 0.53 breaks lintian-brush
- merged portsmf 0.1~svn20101010-6ubuntu1
- synced golang-github-prometheus-exporter-toolkit 0.7.1-3
- synced prometheus 2.33.5+ds1-2, test fails on armhf -> opened bug and forwarded upstream: https://bugs.launchpad.net/prometheus/+bug/1978803
- sponsored https://bugs.launchpad.net/ubuntu/+source/keyutils/+bug/1977982
- synced python-django-formtools 2.3-1
- dcmtk: plastimatch FTBFS Debian #1012439, upstream https://gitlab.com/plastimatch/plastimatch/-/issues/85 -> remove until fixed: https://bugs.launchpad.net/debian/+source/plastimatch/+bug/1978709
- python-rocksdb: FTBFS Debian #1012074 -> Upstream is DD
- balboa: Debian #1012629 -> Missing liblz4-dev build dependency -> uploaded rocksdb 7.2.2-4ubuntu1 with fix
- paperwork for attending DebConf 2022
- Continuing to work on systemd v251 merge
- Getting autopkgtests passing
- Confirmed that upstream changes WRT configuring PCR index in sd-boot have net zero impact for Ubuntu
- Started ubuntu-devel thread about systemd-oomd issues
- Wrote some patches for systemd-oomd and have been testing those, and will start an upstream discussion about those soon.
- Opened a PR for an unrelated systemd-oomd bug I found while testing.
- SRU flash-kernel: sponsoring needed
- flash-kernel kinetic uploaded by @waveform
- BT works: packaging brcm_patchram_plus as it is needed for loading the firmware but have to figure out why the bluez-firmware firmware does not work on this board and how I will ship the correct one to have the right dependencies for this package
- SRU u-boot-nezha: @ginggs uploaded it
- Kernel: v2 of the patch fixing the PFN_MASK
- build for Nezha and VisionFive PR
- with @juliank sorted out why the new lxd remote unit isn’t working, launched another one using focal this time instead of jammy
- updated the autopkgtest-cloud documentation so that it says to use focal for the lxd remotes
- updated https://launchpad.net/auto-package-testing page with a proper url to results
- opened a bug about stray systemd units being leftover when a lxd-remote is removed
- opened a bug regarding the postdeploy script in the mojo spec for autopkgtest-cloud failing
- investigation into repeating systemd (from @enr0n’s PPA) tests, sorted out how to get the removed from the queue as they kept being retested
- due to the above - reported a bug regarding what is and is not considered temporary failures for autopkgtests
- added systemd-upstream/ppc64el to big_packages for bluca from upstream systemd CI testing
- investigated an issue with amd64 tests failing somewhat regularly on 2022-06-15
- submitted an RT regarding networking failures in lgw01 that were discovered due to the above
- looked at the SQLite database of autopkgtest results for potential data mining
- as a result of a flavor lead request I set the “number of rebuilds a product owner can request per day” to 6 via the admin interface
- merged a hint MP regarding firejail for jammy
- modified the apport code location in the Launchpad project for @bdrung
- special Jammy SRU review of network-manager for a regression bug fix
- tested the recommended changes in bug 1977644 regarding a fix for slow booting BIOS systems
- reported a bug to the snapcraft forum about contact information not being available for snaps in the Ubuntu live environment
- reported a bug regarding the attach_journal_errors crashing
- SRU team shift
- completing HR management training
- submitted an RT regarding creating a queue for the Ubuntu QA team
- moved JIRA cards from the Foundations board to the Ubuntu QA one
- Attempted to debug an issue with the rpi-imager snap
- Tested pipewire audio stack on kinetic with the Pi 4, where the wrong audio output is picked by default by switch-on-connect under pulseaudio (LP: #1877194)
- Responded to some more bits of the sbuild merge review (LP: #1973285)
- Experimenting with systemd-oomd settings on small systems (LP: #1972159)
- Firefox performance measurement and work
- Uploaded flash-kernel merge (LP: #1973801)
- Pi meetings
- Reworked the final YAML consolidation PR.
- Reviewed the
- Worked on the current SRUs to Jammy, Impish, Focal and Bionic in prevision of the upcoming security releases
- General bug triaging
- Worked on a spec to formalize tracking vendored dependencies in Rust packages.
- Looked into failing tests for rust-parking-lot, sent a fix to Debian.
- Sponsored a big update (and SRU to Jammy) to
s390-toolson behalf of @fheimes, fixing quite a few bugs: LP: #1971993, LP: #1960119, LP: #1959548, LP: #1971959, LP: #1978323
- Reworked part of
dh_sphinxto use a full JSON parser instead of a hand-rolled solution, thus avoiding the need of the libunicode-escape-perl MIR