Friday 28 June 2024

+1 maintenance report (Jun 24-Jun 28)

Hi there!

You can also read this report on Ubuntu Discourse -> https://discourse.ubuntu.com/t/1-maintenance-report-week-26-2024/46071

I was on my third +1 maintenance this week and this is my shift report.

I began Monday with a run through the update_excuses page and found quite a few autopkgtest runs on arm64 failing due to what seemed like infrastructure issues. I resubmitted all of them. Later I also saw infrastructure failures on armhf, ppc64el and s390x. Re-submissions have had a near 100% success rate. Here's the list [1] of packages that migrated following re-submissions (apologies for not recording the migrating versions). 

I wasn't really planning to target any particular class of packages stuck in -proposed. I went scouting for universe packages stuck on problems that I could locally reproduce, investigate and comprehend. Following is a list of universe packages that I investigated and have attempted to fix. 



==== python3-numpysane ====
A build-time test in C depends on header files from python3-numpy which were earlier symlink'd from `/usr/include/python3.12`.  After the symlinks were removed the test started failing and was initially reported to Debian [2]. A fix [3] was devised upstream. But this fix needs a new builddep `pkg-config` to take effect.
Bug-report - https://bugs.launchpad.net/ubuntu/+source/python-numpysane/+bug/2071467
Merge proposal - https://code.launchpad.net/~pushkarnk/ubuntu/+source/python-numpysane/+git/python-numpysane/+merge/468434



==== razercfg ====
A commit[4] related to the new version (0.43-2) in debian/unstable hard-coded a python3.11 path in the razercfg.install file which causes the dh_install to fail on oracular.
Bug report - https://bugs.launchpad.net/ubuntu/+source/razercfg/+bug/2070171
Merge proposal - https://code.launchpad.net/~pushkarnk/ubuntu/+source/razercfg/+git/razercfg/+merge/468038



==== python-qwt ====
Autopkgtests for python-qwt fail because of a ModuleNotFoundError for 'pytest'. Adding python3-pytest as a test dependency fixes the problem.
Bug report - https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1074168
Merge request - https://salsa.debian.org/science-team/python-qwt/-/merge_requests/1)


==== streamlink ====
Package streamlink has autopkgtest failures because the failing tests depend on python3-pytest 8.2.2-1 which is also stuck in -proposed. I have filed an update_excuse bug which should be closed once both pytest and streamlink migrate.
Bug report - https://bugs.launchpad.net/ubuntu/+source/streamlink/+bug/2070187


==== dkimpy vs. authheaders ====
Fix for an upstream bug report [5] seems to have changed the ARC signatures generated by dkimpy. A test from package authheader (that depends on dkimpy) needed to reflect this change in the ARC signing. When I realized this and visited debian salsa, I saw a new authheader release with this change. This version (0.16.3-1) has now been synced from Debian and dkimpy has migrated today.



==== translate-toolkit ====
The autopkgtests for 3.13.0-2 fail with the python3-aeidon (src: gaupol) 1.15-1.  This is reported [6] by the debian CI as well. It's a ModuleNotFoundError for charset-normalizer, both in build-time tests as well as autopkgtests. Adding python3-charset-normalizer as a builddep and test dependency fixes the problem.
Bug report 1 - https://bugs.launchpad.net/ubuntu/+source/translate-toolkit/+bug/2070381
Bug report 2 - https://bugs.launchpad.net/ubuntu/+source/translate-toolkit/+bug/2070382
Merge proposal - https://code.launchpad.net/~pushkarnk/ubuntu/+source/translate-toolkit/+git/translate-toolkit/+merge/468161



==== libmina-sshd-java ====
This package is stuck in -proposed for 45 days. A Java-based SFTP test fails only on Launchpad builders. I wasn't able to reproduce this issue in my local chroot. I wrote up a bug report and a possible way to work-around the problematic code that could fail in chroot environments.
Bug Report - https://bugs.launchpad.net/ubuntu/+source/libmina-sshd-java/+bug/2071358
Merge Proposal - https://code.launchpad.net/~pushkarnk/ubuntu/+source/libmina-sshd-java/+git/libmina-sshd-java/+merge/468349


==== wiredtiger ====
This package is stuck in -proposed for 66 days. This issue had an existing bug report [7]. A build-time test fails due to a buffer overflow which I have attempted to fix. However, the test PPA builds now fail for different reasons on armhf [8] and riscv64 [9].
Merge proposal - https://code.launchpad.net/~pushkarnk/ubuntu/+source/wiredtiger/+git/wiredtiger/+merge/468239



==== wireshark vs. libxml2 ====
I opened a bug report and the package maintainer reported that the problem had been fixed.
Bug report - https://bugs.launchpad.net/ubuntu/+source/wireshark/+bug/2070447


==== url-normalize ====
This package is stuck in -proposed for 118 days as of today. Build-time tests fail with "ModuleNotFoundError: No module named 'six'" causing FTBFS. Adding python3-six as a build-dep fixes this issue. This issue is not seen on debian/unstable because python3-six is pulled in as a transitive dependency.
of (IIRC) python3-requests.
Bug report - https://bugs.launchpad.net/ubuntu/+source/url-normalize/+bug/2070517
Merge proposal - https://code.launchpad.net/~pushkarnk/ubuntu/+source/url-normalize/+git/url-normalize/+merge/468257


==== cura-engine ====
This package is stuck in -proposed for 87 days as of today. This issue had an existing bug report [10]. It was another case of a buffer overflow crash, but in autopkgtests.
[Merge proposal](https://code.launchpad.net/~pushkarnk/ubuntu/+source/cura-engine/+git/cura-engine/+merge/468332)



==== aevol ====
The package is stuck in -proposed for 191 days as of today. And this turned out to be another buffer overflow crash!
Bug report - https://bugs.launchpad.net/ubuntu/+source/aevol/+bug/2071412
Merge proposal - https://code.launchpad.net/~pushkarnk/ubuntu/+source/aevol/+git/aevol/+merge/468397

And that's it for this week! Thank you for reading through.

Regards,
Pushkar
Posted by at