Comment 23 for bug 2051939

Regarding autopkgtest regressions:

On Mantic:
* the surf tests are very flaky[1], and have been so for a while[2]. With some retriggers, arm64 could pass, but with the current load on the infra, I couldn't make armhf to pass. Still the error reported in the log file[3] is `Too few characters detected (0)`, and is thrown by the surf test itself. Nothing in this test makes calls to autopkgtest provided commands, so I don't expect this regression to be actually caused by this current SRU.

On Jammy:
* gscan2pdf and sbuild were transient failures that retries fixed easily.
* systemd, on the other hand, is directly related to this SRU:
  The systemd testsuite, for some reason, tries to patch the `autopkgtest-build-lxd` here[4], but the patch doesn't apply anymore. Given that this file is not present in more recent versions of systemd, I'm not sure if fixing this test is the right way to spend time.
  As for the `networkd-test.py` test, this one has been observed to be flaky in many other runs with different triggers [5]

[1]: https://autopkgtest.ubuntu.com/packages/s/surf/mantic/armhf
[2]: https://autopkgtest.ubuntu.com/packages/surf/jammy/armhf
[3]: https://objectstorage.prodstack5.canonical.com/swift/v1/AUTH_0f9aae918d5b4744bf7b827671c86842/autopkgtest-mantic/mantic/armhf/s/surf/20240227_144458_eb091@/log.gz
[4]: https://git.launchpad.net/ubuntu/+source/systemd/tree/debian/tests/tests-in-lxd?h=ubuntu/jammy-updates#n12
[5]: See https://autopkgtest.ubuntu.com/results/autopkgtest-jammy/jammy/amd64/s/systemd/20240227_140930_112c3@/log.gz and https://autopkgtest.ubuntu.com/results/autopkgtest-jammy/jammy/amd64/s/systemd/20240227_132149_18a85@/log.gz and https://autopkgtest.ubuntu.com/results/autopkgtest-jammy/jammy/amd64/s/systemd/20240227_051630_ed958@/log.gz for examples.