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.
Did a bunch of working with IS and others to get the Ubuntu metrics Grafana opened up to the public! I’ve written a couple of example metric collectors, as you can see there, and made those into graphs. If you’ve got an idea for something in the Ubuntu community that could do with a dashboard, let us chat. I’ll be writing an email to ubuntu-devel about this later this week.
Grafana here - it’s quite basic now(!) but over time we’ll end up organising it into better dashboards and panels.
Ideas might be like the following. In each case code will need to be written to collect the data you need;
How big are the SRU/sponsorship/unapproved queues? Does your team have review requests outstanding?
What’s the bug situation looking like for your packages?
Are your snaps/debs up to date with their upstreams or Debian?
The arm64/armhf runners are still being unstable in autopkgtest, so I watched those a bit.
Reviewed some fixes from others for autopkgtest, and carried on working on the charm for the new deployment. Talked with IS and some other people about how we should not expose the URLs directly into swift in the future, but proxy them instead. They’ll be something like https://autopkgtest.ubuntu.com/results/autopkgtest-hirsute/hirsute/.../log.gz, which will be stable over time unlike the current ones.
Got together with @juliank today, before Google exploded, and fixed & uploaded update-manager and update-notifier to hirsute and focal. These updates implement the “post install” OEM experience: if you buy a device (e.g. a laptop), install Ubuntu on it yourself, and then Canonical certifies/enables it some time later, update-manager will offer you an update to that experience - which might for example switch you to the OEM kernel track for better hardware support, or install some extra packages. See more info in the SRU bug here.
arm flaky tests - worked with @laney (thanks so much for the brainstorming session!) to find a path forward. Ended up pushing a change to use timeouts to allow the failures to be marked as skipped… works but there’s an alternative approach we could take… a little more investigation is needed by me on this but hope to wrap it up this week before the break.
worked on coinMP bug. There’s a 7.0.3-0ubuntu4~ppa2 new build (built for hirsute) that should have this enabled. Rico included it in the 7.0.4~rc2-0ubuntu1 build but turns out many MIRs are needed. Imho, these MIRs are not a priority this week but I can make a trello card to complete for next year.
worked on LO snap actions issue. Currently working out kinks in override-build logic to properly subtitute actions.
fixed the quadrapassel snap failing to build on some archs
replying to some review comments about the thunderbird/default/snap fix
sponsored an ibus fix from Gunnar
debugged xdg-settings url-scheme-handler interactions not working in the thunderbird snap, turned out that snapd got support for that feature but the core wrapper didn’t get updated to use it. Submitted patches to fix the issue
reviewed my notes on how to do an ISO building locally and cleaned a bit since some fixes got merged.
discussed the 78 update in focal-proposed being blocked on making tinyjsd uninstallable with the SRU team − I am going to update the SRU bug with a justification of the breaks and will work with the SRU team to find an acceptable solution
chromium
updated stable to 87.0.4280.88
updating beta to 88.0.4324.41
updated dev to 89.0.4350.4
started experimenting with building the snap against the gnome-3-34 snapcraft extension
PostScript Printer Application: Continued the development and the testing. Investigated the possibility to poll default option settings from the PostScript printer. One can poll the presence of installable accessories if the PPD provides appropriate code, but one cannot poll which media are loaded in each tray. I have a web admin page for (manually) setting the accessory configuration now. Bought a second paper tray for my printer to test this.
PAPPL: Now with PAPPL having all features needed for the PostScript Printer Application (PAPPL issues: open, closed) Michael Sweet has released version 1.0.0.
CUPS Snap: Still waiting for @jamesh’s API extension of snapd to pass Canonical’s security team review so that the CUPS Snap makes it into the Snap Store.
sane-airscan: Waiting for the MIR to get reviewed by the security team.
I updated the build scripts to build the a customised snapd that enables the user-daemons and dbus-activation experimental features. This was enough to get the image correctly seeding again on first boot. The custom snapd gets replaced with the standard one on snap refresh, but we’ve finished seeding at that point and have the experimental features enabled via the gadget snap.
With the strict confined desktop session updated to support activation of the org.gnome.Terminal service, I was able to get gnome-terminal running from the confined session. That demonstrates both user daemons and dbus activation functioning.
snapd work:
I had a chat with Alex about security reviews of snapd PRs (including the one @till-kamppeter is waiting on). He had been busy with other things, but has added them back to his todo list.
I collated some feedback on the notification code in snapd, as is shown when the refresh awareness experimental feature is enabled. I put together a couple of PRs laying ground-work for improving the notifications:
snapd PR #9772, which changes the testing strategy from mocking the D-Bus connection to using a private session bus and fake notification server.
snapd PR #9779 (merged), which changes the private session bus fixture to disable service activation. This ensures that D-Bus services installed on the test system are not accidentally activated by the test suite, and makes it possible to reliably test for behaviour in the absence of a service.