The previous status is here: Foundations Team Updates - Thursday 26 January 2023
- pygments migration
- MIR: review log4cplus, https://pad.lv/2003549
- MIR: Discuss libheif + dependencies with @vpa1977, https://pad.lv/1827442
- MIR: investigate and postpone new dnspython dependency MIRs, https://pad.lv/2004272
- MIR: review process PR: https://github.com/canonical/ubuntu-mir/pull/4
- MIR: update status for editorconfig-core, https://pad.lv/1984104 (thanks @vorlon for promotion o/)
- MIR: trying to fix tk8.6 false positive component-mismatch upstream: https://bugs.debian.org/1030148
- inclusive naming ++, PR#303
- Upstream bugs started a good discussion:
- Review, merge and implement PRs:
- review & merge memory leak: https://github.com/canonical/netplan/pull/321
- review & fix Permanent MAC: https://github.com/canonical/netplan/pull/278
- review duplicate routes: https://github.com/canonical/netplan/pull/320
- implement netplan_delete_connection API: https://github.com/canonical/netplan/pull/322
- review & fix autopkgtest stability: https://github.com/canonical/netplan/pull/223
- review vxlan tristate: https://github.com/canonical/netplan/pull/311
- some compliance training
- job posting review
- several take-home test reviews
- mid-cycle sprint sessions
- review ubuntu-image, PR#111
- raspi-config spec review
- Added a wheelhouse.txt file pinning pip versions thereby fixing the charmcraft pack failure with autopkgtest-cloud-worker. Added the same wheelhouse.txt file for the autopkgtest-web charm.
- Updated, tested the autopkgtest-cloud-worker charm in staging. Then updated the charm in production.
- Wrote, tested and added an lpcraft configuration file to the ~ubuntu-release branch of autopkgtest which runs pyflakes. This’ll get run on every MP for that branch.
- Provisioned 2 new armhf workers replacing two downed units. Additionally, used the
same_hosthint to make the node creation process less painful.
- Submitted an MP updating our documentation so that the
same_hosthint is documented.
- Wrote, tested and added an lpcraft configuration file to the ~ubuntu-release branch of autodep8 which runs autopkgtest against the branch (which contains autopkgtest tests).
- Submitted a merge proposal which improves on the above by using autopkgtest from the ~ubuntu-release branch of autopkgtest in addition to the packaged version for Focal.
- Updated, tesed the autopkgtest-web charm in staging. Then updated the charm in production.
- Reviewed, merged an MP from Simon Quigley updating the ubuntu-release branch of autodep8. Then updated the version of autodep8 in staging and production.
- Removed crash reports from the Error Tracker for a user.
- Added r-cran-prophet, seqan3, psychopy, rust-ahash, golang-github-aws-aws-sdk-go-v2, pycuda, pandas, bmtk, praat to big_packages.
- Uploaded lsb-release-minimal with some test ordering changes as having busybox as the shell broke log file grabbing see bug 2004500 for more details.
- Roadmap sprint week.
- Administrative (travel) work.
- sourceful NEW processing, got the queue down to zero
- processed various removal requests, working towards unblock of
process-removalsfrom Debian; almost ready to remove python2 from lunar, waiting on the kernel packages’ test suite to be updated to python3
- pushing on the python3-defaults transition. Getting closer, but britney now telling us about blockers that it should’ve told us about before (
Roadmap mid-cycle; lots of good discussions
- Made some updates and merged PR to clean up /run
- Made some updates and merged PR to clean up base tarballs
- Began work on spec to support revision assertions in ubuntu-image core builds
- Investigated “bug” for Field Engineering team. This turned out to be an intentional design decision from snapd
- Meetings to demo ubuntu-image and support the CPC gadget effort
- Finishing up +1 maintenance, see ubuntu-devel mailing list for more details
- Candidate interviews
- Take home test reviews
- Attended roadmap sessions
- proposed-migration: reprepo vs libarchive
- Truly a weird issue. It gets a broken pipe on s390x only, but consistently. I’ve tried a few different sets of triggers but had no luck.
- Sync’d dipy and nbconvert from Debian
- Uploaded fixes for the following packages:
- Reviewed and sponsored another dotnet7 kinetic SRU for @mirespace
- numpy migrated!
- Added migration hints for psi4, mailman-hyperkitty, mailman3 and siconos
- Added spyder, python-bumps and python-pymbar to big_packages
- More autopkgtest retrying to help Python 3.11 / proposed-migration along
- Requested removal of the following packages from lunar:
- pairtools s390x binaries
- theano, deepnano and keras
- nipype and heudiconv
- Investigated a new autopkgtest failure that showed up in libcap2 that happens only on armhf that seems to be related to the test environment. We are considering disabling that test on armhf for now as it’s not reproducible in Lunar.
- Investigated autopkgtest failures on the new version of python-typing-extensions package
- Addressed some issues found in the MIRs I’ve submitted recently
- Finished implementing the first batch of autopkgtests for the NetworkManager+Netplan integration
- Added a new step to our Github CI to check for runtime memory issues in the Netplan C code (PR#321)
- Finishing a fix for some vxlan options that couldn’t be disabled (PR#311)
- Working on a fix for an issue that causes Netplan to ignore some routes when they are defined in different files (PR#320)
- Attending multiple meetings from the mid-cycle sprint
- travel planning
- hiring packet review
- casper - LP: #1986781 - verify SRU for Jammy
- PR: 1546 - review recent work on main and bring over two fixes for the stable branch, a CI fix and a logging fix that was confusing users.
- PR: 1545 - after LP: #1986781 was implemented, systemd service status started to be printed on top of the TUI (LP: #2004040). Send to systemd signal SIGRTMIN+21, which is documented as a method to set show_status=0. Also merging to kinetic (PR: 1548)
- PR: 1541 - tweak strings shown in TPM FDE cases
- LP: #1993633 - UI portion can be reproduced in dryrun, but why is probert sending weird data?
- PR: 5 - fix license, and the resulting v 0.1.1 packaged accepted to archive
- proposed-migration: openstack-trove - investigate and write up LP: #2003974
- merged PR 1537 to get the basics of mirror testing.
- fixed regressions caused by absence of pkg_resources in the subiquity snap (PR 1542)
- cherry-picked Ubuntu Pro patches for the upcoming 23.02.1 subiquity release (PR 1543)
- opened and merged PR 1544 - various fixes for mirror testing
- opened PR 1547 to get output of mirror testing translated in the user’s selected language.
- working on having automatic fallback to a different mirror.
- on +1 rotation this week. Thanks @schopin for the baby-sitting
- uploaded ubiquity 23.04.4 to fix test-suite when run with Python 3.11 (thanks @vorlon for sponsoring!)
- re-triggerred FTBFS python-leidenalg build against updated python-igraph
- asked for a bunch of no-change rebuilds for small transitions
- uploaded fix for kitty FTBFS on ppc64el (thanks @tsimonq2 for sponsoring!)
- uploaded fix for django-rich FTBFS when TERM=unknown: LP: #2004553. Fix was forwarded upstream.
- merged git-annex but it fails to link in Ubuntu because of a relocation issue. Not sure how to move forward.
- looking at mailman3 autopkgtest regressions.
- python3-defaults migration is blocked by metalfinder: Retry with metalfinder 2.1.6-2 (which disables flaky tests) worked
- Assert that Python version is PEP440 compliant and handle more Debian versions correctly in
- Add type annotations, check them with mypy, fix issues found by mypy
- Upload distro-info 1.4 to Debian unstable
- Assert that Python version is PEP440 compliant and handle more Debian versions correctly in
- Make Python version PEP440 compliant (got merged immediately)
- Uploaded python-debian 0.1.49ubuntu2 to lunar with the fix
- Make Python version PEP440 compliant
- Uploaded reportbug 11.6.0ubuntu2 to lunar with the fix
- Make Debian Python version PEP440 compliant
- Uploaded duecredit 0.9.1-1ubuntu1 to lunar with the fix
- Use PEP440 compliant version in setup.py (LP: #1991606)
- Fix issues found by flake8 on the Python scripts
- Check Python scripts with flake8 again
- Format Python code with black and run black during package build
- Sort Python imports with isort and run isort during package build
- Replace deprecated optparse with argparse
- requestbackport: Remove useless loop from locate_package
- reverse-depends: Restore field titles format
- Fix all errors and warnings found by pylint and implement most refactorings and conventions. Run pylint during package build again.
- Uploaded ubuntu-dev-tools 0.192 to Debian unstable
- ci: Run woke on the source code
- refactor: Simplify cmp() in CrashDatabase._duplicate_search_signature
- tests: Expect crash .class and .jar to be present
- test: Fix test_deleted_working_directory for pytest-cov 4.0
- refactor: Introduce get_process_environ
- fix(whoopsie-upload-all): Use NoninteractiveHookUI for add_hooks_info
- Converted ubuntu-packaging-guide to git for @vorlon: https://code.launchpad.net/~ubuntu-packaging-guide-team/ubuntu-packaging-guide/+git/main
- Bit of merge request merging
- Finished up the specification for archive snapshot support in apt
- In free time on Saturday, quickly hacked up a
apt add-sourcescommand that (downloads and) installs a deb822 .sources file into sources.list.d.
- Respun shim-signed with fixes for arm64
- Cherry-picked grub2-unsigned 2.06-2ubuntu16 to stable series
- Wrote a draft specification for secure boot packaging
- Writing a secure boot package workflow tool
- Wrote blog post about the secure boot key rotation https://blog.jak-linux.org/2023/02/01/ubuntu-key-rotation/
- Got my FOSDEM first leg train seat reservation which wasn’t bookable during initial booking
- Responded to review of spec for Ubuntu flavour of raspi-config (need to respond to second round of comments)
- Investigated Firefox acceleration on Pi (requires mesa currently in jammy, and for base of snap to move to Core 22; current issues with building Firefox on Core 22)
- Proposed migration of
pybuild-autopkgtestwasn’t finding the
python3-exceptiongroupbuild-dep (despite it being explicitly listed in
d/control, as opposed to the python-version limited requirement in
setup.py); possibly related to LP: #2002819 (thanks @schopin for noting that!) or they may be some other issue (still digging into the tooling). Fixed by @tsimonq2 in the meantime, by-passing the autodep8 tests
- Verified SRU of
- Uploaded new version of
- Uploaded new version of
rpi-eepromfor lunar (LP: #2004569); this may become an SRU in future
- Investigated slack space at the end of our images (rpi-imager#480); there’s only about 200MB slack at the end (in line with fudge-factor in livecd-rootfs), but the rootfs is only 66% full (1.2GB free!)? Need to dig further…
- 22.04.2 release meeting
- Interviews & tests
This week has been mostly about the upcoming glibc 2.37 upload:
- Posted a heads-up announcement on firstname.lastname@example.org and Discourse
- Investigated and reported a regression in strncat on modern amd64 hardware (fixed upstream)
- Prepared a PPA upload of the upstream tarball
Still working on re-bootstrapping rustc 1.62 for the kernel in Lunar. The final round of builds is ongoing, finally
Jammy Point Release
- test images
- SRU testing for U-Boot in Focal
Lenovo X13s installer image
- evaluate different approaches to create a makeshift installer
- Lost my LUKS passphrase and therefore tried to recover my data (still trying) and reinstalled, reconfigured, …
- Continued on the crypto configuration specification which is reaching completion after answering comments
- Finished proposed migration of
astroplan. It took longer than expected due to some issues with
- Currently working on updating Rust and Cargo to 1.66.1 and 0.67.1 respectively. Expect this to take a while.
- Currently investigating the possibility of creating a minified Ubuntu image for Rust application deployment.
Short week due to public holiday
- Written MIR reports for libheif and dependencies https://pad.lv/1827442
- working on minimal set of files for Java 8 runtime. Proposed set