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.
cups-filters: Fixed several bugs in time for Final Freeze of Eoan, due to tests by the reporters of the bugs and discovery of further bugs released 1.25.8, 1.25.9, 1.25.10, and 1.25.11 upstream and uploaded them to Eoan. What got fixed were several problems with remote IPP printers using the old IPP 1.1 standard and not the current 2.x (Upstream bug report), an incompatibility with Ghostscript 9.27 (Upstream bug report), and a crasher caused by missing page size names in CUPS’ list of known page sizes (Ubuntu bug #1847488).
Google Code-In 2019: Application period for mentoring organizations started. Studied the possibilities whether to participate and started the application for the Linux Foundation.
fftw3 ftbfs in eoan (armhf only): eoan: More days spent experimenting with build options to narrow down the problem. On hold now until my new Pi arrives this week to speed up the debugging process.
bionic: Built an Nvidia+bionic machine just for this bug. Not reproducible on bionic with any driver version so far. There were also no bug reports from bionic systems.
I worked out a way to fix the test failure in glib MR!1111: essentially switch from g_test_dbus_down to a g_test_dbus_stop call. Since the code keeps a reference in to the bus in a global, we can’t use an API that verifies the reference count drops to zero.
Improved snap support in xdg-desktop-portal:
As described in my forum post, I put together PR #7588 that implements a snap internal portal-info $pid command that returns the snap name, desktop file ID and whether the snap has network connectivity. This
There was a request to split out the code to register snap internal sub-commands. That was fairly easy to separate since it was just the first commit, so I created PR #7589. This one apparently needs review from Samuele, but it has support from some other snapd core developers.
snapd session agent:
There is in-principle approval for the systemd API. Since the session agent doesn’t run on the distros with ancient systemd versions, it looks like I can probably remove some of the complexity from the “stop” action compared to how system level units are handled.
Implement initial grpc separation and binds to an Unix socket
Implements a logger, so that daemon can have its own logging level and client another one. The client can get logs streamed for its own request while processing it, even if the daemon is at a lower debugging level.
Implements code generator to extends generated protobuf message to embed log generation.
On the client side, wrote a GRPC interceptor at request creation time for the same transparency level
Make zsys 20.04 only compatible with golang 1.13 to fix CI (errors vs x/errors package changes)
Moved from “org.zsys” to “com.ubuntu.zsys” domain (zsys, grub2, grubzfs-testsuite, grub2-signed, zfs-linux & ubiquity). Consequently, note that only new install starting with ubiquity 19.10.19 will be zsys compatible. People with existing installation can annotate manually, replacing with “zfs set” tags their datasets from org.zsys to com.ubuntu.zsys.
Changed (with jibel) how zfs is installed to limit it to zfs only selection. Changes the seeds + meta to reflect that and coordinate with flavors which opted-in.
Feedback online (email with Richard, discourse, twitter, ars technica…)
Prepared blog post (part 2) for ZFS on 19.10, to be published this week.
a user pointed out that the chromium snap CSD bug is fixed with gtk-common-themes from candidate, I confirmed their observation and Ken promoted it to stable