Previous week’s status can be found at: Foundations Team Updates - Thursday 19 Oct 2023 - #17 by vpa1977
- Reassign dual-stack routing issue to cloud-init, https://bugs.debian.org/1052535 (cc @falcojr)
- MIR review: Bug #2038929 “[MIR] libnet-idn-encode-perl (as libmail-dmarc-per...” : Bugs : libnet-idn-encode-perl package : Ubuntu
- review & sponsor & verification of NetworkManager YAML delete SRU, https://pad.lv/2040153
- prepare NetworkManager notify-reboot SRU, https://pad.lv/2040292
- review NetworkManager arm64 autopkgtest fixes, https://pad.lv/2023183
- prepare stable-proposed-updates fix for Netplan’s autopkgtests vs systemd v252.15 in Debian Bookworm
- Investigation of NM-Netplan bug, https://pad.lv/2040153
- Netplan ecosystem summary
- Debian devref discussions, https://salsa.debian.org/debian/debian-reference/-/merge_requests/14
- “Netplan everywhere” docs & blog post, https://github.com/canonical/netplan/pull/418 (cc @danilogondolfo)
- review & sponsor Mantic wireguard & EAP SRU
- Ubuntu Core 20 /snap/bin/nmcli SRU, https://pad.lv/1959570 (cc @alfonsosanchezbeato)
- 0.104-0ubuntu2~20.04.3 can be dropped from the UNAPPROVED queue
- 0.104-0ubuntu2~20.04.4 should address all remarks and be accepted into focal-proposed (cc @sil2100)
- plenty of “Greenhouse gardening” (written interview, take-home tests)
- roadmap massaging
- Meeting with Tim
- PR #150 - Support the keep-enabled parameter in ubuntu-image extra-ppas
- PR #148 - ubuntu-image ‘pack’ support
- PR # 155 - Clean image build leftovers
- PR #143 - The ‘fstab’ of ‘customization’ destroy original fstab content
- improved and merged
- PR # 157 - Use eatmydata to speedup image build
- waiting for a review
- PR # 159 - Always look for the gadget.yaml file in the same place
- soon to be merged
- PR # 160 - Distinguish build and target components
- let users define what components/pocket should be used at build time and in the resulting image
- waiting for a review
- PR # 158 - Check dependencies vulnerabilities
- use govulncheck in CI to identify CVE in our dependencies
- build with golang 1.21
- waiting for a review
- PR # 161 - Improve tests
- The whole test suite was taking around 40min to run, for a 91% coverage
- Mark some tests as “long”. Short tests run in ~6s, with a 84% coverage. So we can iterate a lot faster with a decent confidence.
- Do some cleaning/harmonization on tests.
- Mark some tests as able to run in parallel.
- Add a “basic” chroot, generated only once (if needed) and copied by tests needing it (avoid doing the same chroot generation multiple times)
- waiting for a review
- Some more research on alternatives to ubuntu-image.
- Some ubuntu-image bug grooming
- Reviewed boot loader testing framework
- Some work on APT snapshot SRU verification
- Hiring stuff
- Packaged python-uefivars for setting up secure boot keys
- First version of PC BIOS, and UEFI AMD64/ARM64 SB tests working
- Started working on the ability for UKIs to contain multiple DTBs.
- rdma-core: release new upstream release 48.0-1
- Add noble: https://salsa.debian.org/debian/distro-info-data/-/commit/e510c2cf504a7ba3ca91e42ff7e3add2e4bee6c1
- After noticing differences between distro-info-data and Wikipedia, I checked the dates for Debian and Ubuntu. I corrected Wikipedia and distro-info-data and references the relevant mailing list or Debian release announcements.
- setuptools: Fixed Aptdaemon crashes when there are any non-PEP-440 packages on the system due to setuptools 66.x.x for Ubuntu 23.04
- dhcpcd: 1:10.0.4-1 got uploaded to Debian unstable with all our changes. So we can sync that version after the archive opening.
- dracut: Two merge request for Debian were merged. So I prepared another one to remove the Ubuntu delta: Address some lintian complaints
- Helped debugging When upgrading to Mantic, it fails to install snap firmware-updater
- Reviewed fix: MachineType fstring
- hookutils: skip checking stat for skipped DMI files and add unit test for apport.hookutils.attach_dmi
- feat: remove support for Upstart
- setup: Do not fail when pkg-config is missing
- style: merge consecutive f-strings
- test: remove dead system test code in test_python_crashes
- ci: upgrade to actions/checkout@v4
- launchpad: catch import error if httplib2 is missing and move tests into tests/integration/
- Correct get_id_url() for testing Launchpad instances and use launchpadlib.uris.lookup_web_root
- test/launchpad: Some refactorings
Remaining bits from the +1 maintenance week:
I spend my time ensuring that all blocked packages failures are documented and reported in Debian / upstream where applicable. I created a bunch of bug reports:
- rmlint 2.9.0-2.5 fails to build: test_stdin.test_stdin_read hangs
- aircrack-ng 1:1.7+git20230807.4bf83f1a-1: FTBFS due to failing test-wpapsk-cmac
I ended on cross-toolchain-base-mipsen which is 71 days old as of this Friday’s report (i.e. in the queue since 2023-08-10). From the 103 packages in the queue, the oldest 58 packages have a update-excuse bug linked and 8 of the more recent ones. This leaves 37 packages that still need to be looked at.
pyina and pathos cannot be built because they need multiprocess >= 0.70.15, see Update to newer upstream (>= 0.70.15) to allow python3-pathos to build. So I went ahead and updated this Python modules using my Debian Python Team hat.
I created the update-excuse bug #2039968 and linked the already existing bug in Debian. Since it is a Python package under the Debian Python Team, I tried to fix it. The
cmake configuration file has options to disable ssse3 and aes, but it hard-codes
-mstackrealign. Then the build fails due to the missing
wmmintrin.h header. So I decided to report the issue upstream to let upstream fix it: https://github.com/Legrandin/pycryptodome/issues/774
Following package should be (auto-)synced after the archive openening:
- Sync multiprocess 0.70.15-1 to allow building pyina and pathos
I’ll be away next week so I’m going to include what I think will be done/uploaded in the coming days
- openssl SRU for Jammy: Bug #2033422 “openssl: backport to jammy “clear method store / q...” : Jammy (22.04) : Bugs : openssl package : Ubuntu
- one performance fix for concurrency
- fixes for “ignored SMIME signature errors” and “imbca engine dumps core” (see link above)
- dropping a fix for blowfish using the wrong default key size because it would introduce intra-jammy incompatibilities upon a simple upgrade
- gnutls: forcibly disable TLS 1.0 and 1.1 which we meant to disable system-wide but were still typically enabled
- openssl: remove dead code in postinst and remove use of debconf which was useless but unneeded and triggering an old and mysterious issue with debconf locking
- crypto-config: specification and lightning talk updates
- Preparing SRU to backport golang-1.21 to Lunar/Jammy/Focal as requested by CPC team. LP: #2040269
- Start to drafting policy and process for updating Go toolchain in stable release.
- PR#1851 - some updates to subiquity apport report information in line with apport hook updates.
- PR#1855 reverts the fix for some late command issues with the snap environment (e.g., LP#2032961) since this seems to have just introduced more of them but in the other direction. We will have to revist this in the future with a better approach.
- MP#454563 (in review) - rewrite of apport hook for subiquity to update collected information.
- Doing some work to generalize some package hooks for subiquity, cloud-init,
- PR#226 tiny fix for an f-string typo I found in my apport testing
- Worked with the team for the new series opening (https://autopkgtest-cloud.readthedocs.io/en/latest/administration.html#opening-up-a-new-series)
- MP reviews
- Uploaded new version of livecd-rootfs
- Bug triage rotation
- Investigating interruption of data series for the boot speed measurements
- Reviewed and approved packaging of a new upstream version on Kea
- This happened in Debian: we’re keeping Kea mostly a sync.
- Patch pilot rotation
- Cleared Greenhouse backlog (technical exercises and written interviews)
Had the day off Tuesday. Short week this week.
- MP amending an issue we had with trusty and xenial service files in infrastructure
- working on an MP which introduces a feature which allows users to stop currently running tests from the webUI
- Made an MP in launchpad-mojo-specs to amend an issue we were having with CI and unit tests
- Assisting in seeding new release and debugging issues with new
- Testing branch ubuntu/5.20 in staging with a number of different tests. It’s a new branch prepped a little while ago by Paride which is more up to date.
- Made an MP which introduces upgrade tests to N
- help move doc PRs along with reviews and a CI fix on PR: #1845 + others
- setting up stable build for some fixes:
- LP: #2040654 - regression in late-commands handling - we’re reverting the change that introduced the regression, and evaulating from there
- LP: #2040190 - a chain of problems tracing back to the kernel, as listed in LP: #2034705, results in efibootmgr failing. There appears to be a environment variable workaround, but the first test of the change has failed, so this workaround is still pending.
- patch pilot
- LP: #2036873 - after review, I agreed with Paride’s analysis, did the small tweaks Paride asked for, and uploaded the bootstrap builds for focal/jammy.
- LP: #2039736 - analyzed the request and agreed with submitter that a syncrequest was appropriate to drop the ubuntu delta, which only set a minimum version on python3-urllib, which would not be possible to see unless building unmodified for pre-jammy. But it’s requested already, so no action other than unsubscribe sponsors. Updated bug state to match.
- LP: #2039611 - adjust for noble instead of mantic and upload.
- 24.04 planning
- Setup the prod-proposed-migration environment to run tests for Noble.
seed-new-releasewhich copies forward test results from mantic to noble. However, it needed very little hand holding due to changes we’d made to the script.
- Did some handholding of download-all-results as it timed out multiple times on the autopkgtest-web servers.
- Reviewed an MP from @andersson123 regarding running trusty autopkgtests.
- Submitted an RT regarding adding noble images to the various clouds. Then communicated with the bootstack team regarding adding the same images to PS5 (lcy02).
- Worked with @vorlon regarding backing up the server on which this service runs.
- Helped with the release of Ubuntu 23.10.1.
- Investigation into failure to install firmware-updater snap during a distribution upgrade bug 2039268
- Modified the phased-updater code to no longer consider bionic as it is ESM.
- Merged @waveform’s MP adding devices to raspi/os_list_imagingutility_ubuntu.json. Then updated the branch on changelogs.ubuntu.com.
- Updated release checklist templates with additional information regarding new release opening and release EoL.
- Uploaded distro-info-data adding Noble which “fixes” bug 2040193.
- SRU team duties.
- Reviewing candidate applications.
- Conducted an interview.
- prepared October security release and uploaded to ppa, still building due to the issues with RISC-V builders
- working on adding test suite package to openjdk LP: #2016437
- openjdk-21: bootstrap builds for jammy and focal are in the NEW queue.
- prepared SRU bug for focal LP: #1930541
Now I can login again … probably about time I post this:
ubuntu-imagechanges for cleaning of files from images (LP: #2039434)
- Proposed adding device tags to the rpi-imager metadata (for the forthcoming rpi-imager 1.8.1, LP: #2039887)
- Posted first part of the NBD blog series
- Worked on the split of ubuntu-settings for noble
- Wrote some acceptance criteria for 24.04 roadmap items
- Take home tests