Hi everyone, below you will find the updates from the Desktop team from the last week.
If you’re interested in discussing a topic please start a thread in the Desktop area of Discourse.
We also have our weekly meeting on IRC. We meet on Tuesday at 13:30 UTC in #ubuntu-desktop on Freenode. There will be an “Any Other Business” section at the end where you are welcome to raise topics. These topics might be discussed during the meeting, or afterwards depending on the time, depth of conversation, topic and so on.
bionic: Not reproducible on bionic with any driver version so far. There were also no bug reports from bionic systems. Waiting for someone to confirm, otherwise no bionic fix would be needed or verifiable.
with upstream’s help, figured out a fix for the regression where external hunspell dictionaries were not being loaded, and uploaded to focal
filed bug #1849162 to track the inclusion of lightning in the thunderbird package
chromium
updated stable to 77.0.3865.120
read and answered James’s questions on the discussion about exposing desktop files created by snaps to the DE
confirmed and fixed bug #1848080 (missing --temp-profile option)
confirmed and fixed bug #1848083 (core dump running chromium --help)
wrote on the snapcraft forum to document the requirements to ship man and local man pages in a strictly confined snap
tested whether the chromium-browser transitional package in eoan could be made “architecture: all”: the upgrade path works but on unsupported architectures (ppc64el and s390x) a new installation times out after 30min and fails − not a good user experience
opened and closed the stable/ubuntu-20.04 channel, and invited Ken to collaborate on the snap
updated beta to 78.0.3904.63
updating dev to 79.0.3941.4 (there are Widevine changes that require reworking the patches)
I’m not sure if racy is the right way to describe the problems that resulted or if we just broke some assumptions (deleting /etc/machine-id makes systemd assume that it’s running in first boot mode which might be relevant). It’d be good to understand exactly why things broke the way they did.
Also note that the zero-length machine ID (the initial state) is correct to have. We were working around some kind of kernel bug in the first place.
Did pair programming with xnox that resulted in a fix to ubiquity to fix some very specific breakage (packages with multiarch relationships on the ISO that weren’t completely satisfied on the ISO i.e. nvidia was missing its :i386 recommends)
That was belt and braces for a fix to debian-cd (and subsequent commits) to manually (!!! the real fix is in germinate and other places to support cross-arch relationships) pull in the needed i386 packages for nvidia onto the ISO.
Helped out with archive opening - given that we got a name on time we could start opening pretty much immediately.
autopkgtest:
Some massaging required today; the cloud was not very stable due to the massive load caused by the perl upload.
Worked a bit on the new staging environment. Required some firewall changes in the backend. One region is still broken but I’ve got a ticket with IS to get that fixed.
Google Summer of Code 2019: Attended the Mentor Summit in Munich.Met with many other mentors and org admins, exchanged experience, presented OpenPrinting’s success with GSoC in a lightning talk and held a discussion session about the future work of OpenPrinting, received invitation to EuroBSDCon conference in Vienna in Sep 2020.
Google Code-In 2019: Completed the mentoring org application, lined up sample tasks, proofreading.
Catch up with e-mail after vacation and GSoC Mentor Summit
cups-filters: Checked progress on debugging of cups-browsed for legacy IPP/CUPS printers (Upstream issues #124, #163, and #167).
Update our gsettings override to drop the compromised sks server (following upstream gcr)
Backported a fix from Robert for the background settings segfautling
Sponsored again the pulseaudio/bionic SRU (with fixed changelog to remove extra bug reference)
Debugged a bit a tracker segfault with upstream
Reuploaded the unity-control-center version which was made to use their own schemas, that was supposed to have landed but the billeto copy failed without error
Errors and launchpad bugs triaging, looked at the post release feedback online