Desktop Team Updates - Monday 30th March 2020

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.

Last week’s notes are here: Desktop Team Updates - Monday 23rd March 2020

  • Released gnome-control-center 3.36.1, 3.34.5
  • Did FFE paperwork for fractional scaling UI changes.
  • Investigating snap-store issues with debs.
  • Sponsored simple-scan improvements.

:desktop_computer: Gnome Shell:

:house: Housekeeping:

:chart_with_downwards_trend: Backlog tracking

  • Zsys:
    • Fixed really fixed state removal and garbage collection.
    • Lot of integration tests.
    • Manual functional tests
    • Bug fixing (but not in removal and GC code)
    • Release of 0.4.2
    • Fix import of external pools (LP: #1850130)
    • Fixed in ubiquity and zfsutils to improve boot performance and support for multi disks setups.
    • Bug triage and investigation.
  • Ubiquity:
    • Reviewed and merged code for OEM and Kubuntu.
1 Like
  • Zsys:
    • Fixed really fixed state removal and garbage collection.
    • Lot of integration tests.
    • Manual functional tests
    • Bug fixing (but not in removal and GC code)
    • Release of 0.4.2
    • Fix import of external pools ([LP: #1850130](
    • Fixed in ubiquity and zfsutils to improve boot performance and support for multi disks setups.
    • Bug triage and investigation.
  • Misc:
    • Final review of game mode MIR ack
1 Like

firefox24 firefox

thunderbird24 thunderbird

  • updated to 68.6.0 in focal and backported to eoan and bionic (pending validation and publication by the security team)

chromium22 chromium

snapcraft24 snaps

  • filed a snapcraft bug for running several builds of the same snap in parallel

package24 other

  • Ubiquity:
    • Had some discussions with mpt and uploaded a version which installs hardware-enablement packages automatically without a UI to opt out. Pending further design, we might offer something inside the session to toggle this.
    • Fixed no background in the “install” session
    • Fixed no g-s-d (bad fonts etc) in the “install” session
    • Got designs and worked a bit on the RST / Bitlocker UI, coming soon
  • Some sponsoring of mozjs / gjs / …
  • Got the staging appstream generator working properly
1 Like
  • worked on updating the LO snap to use the gnome-3-34 extension. Not finished yet.
  • updated the LO snap to 6.4.2
  • rebuilt LO 6.4.2 deb and @oSoMoN uploaded it for me so thanks! (the LO deb is now being updated to 0ubuntu3 build though to fix iso builds)
  • usn refreshes: cherrytree, glimpse
  • setup microsoft teams and poked around to prepare for snapcraft summit
  • testing snapcraft in candidate (has gnome-3-34 extension)
  • opened a MR to add gcc to gnome-3-34 extension to help @kenvandine build firefox with the gnome-3-34 snapcraft extension
  • testing various versions of LO to try and find where libreoffice help failure started. Seems to be specific to 6.4.X in ubuntu only. so now i’m looking into differences between our 6.4.2 and debian’s for clues.
  • LAS: wrote proposal to co-locate with SFScon in Italy in November for GNOME and KDE boards to review
  • Focal
    • GNOME 3.36.1 updates
    • Fixed gamemode being active at login, it should only be dbus activated (FFe review comment)
    • Packaged pydrive, needs by deja-for gdrive backups support
    • Reactivated the google drive option in deja-dup
    • Remove the reminna gdm session, we don’t want that one listed by default
    • Updated poppler to the 0.86.1 version, including usual soname change
  • Plymouth
    • debugged spinner theme issues (livecd, fsck, …)
    • Updated the spinner theme to use the Ubuntu font an orange coloring
    • spinner now depends on the label package, should help with livecd
  • Proposed migrations
    • fixed some autopkgtest i386/crossbuild issues
    • handled the new poppler soname transition
  • SRU
    • SRUed an evolution/bionic fix for HTML rendering issue with new webkitgtk
  • Sponsoring
    • libmtp for Till (fixing some printers)
    • console-setup lithuania keyboard layout fixes
    • synced newer hwdata, gedit-plugins
    • plymouth logo fix to be centered again (thanks Daniel)
  • Snapcraft: Regarding the interface through which snapped apps can print via the CUPS snap had a video meeting with @ijohnson, @jdstrand, and Samuele Pedroni. Also discussed with @jamesh in the thread after the meeting. @jdstrand posted a summary. Question is now whether to implement the interface like the one of PulseAudio or with PolicyKit.
  • libmtp: My patch for the UDEV rules to make USB printers not get “audio” group ownership got applied upstream, in Debian, and also the Ubuntu package got uploaded (thanks, @seb128).
  • cups-filters: More thoughts about the further development towards Printer Applications. As CUPS will drop PPD support soon I am thinking about moving all PPD functionality of libcups into a new libppd, to allow for legacy driver conversion into Printer Applications with a minimum need of new code.
  • ghostscript: Re-introduced the -O3 exception for ppc64el again, as the upstream fix of gcc got withdrawn due to a regression (see bug 1862053).
  • SANE: Made the two maintainers of the “escl” and “airscan” join their projects!! (see thread on GitLab) Also did a lot of testing for WSD (Microsoft’s Web Services for Devices) support by the “airsane” driver on my HP OfficeJet 8730 Pro. With my feedback all the problems got fixed and my device is now perfectly scanning with both eSCL (Apple AirScan) and WSD and probably 100s of other multi-function devices, too.
  • PAPPL: The Printer Application Framework from Michael Sweet is progressing. There is a test frontend now and one can enter the web interface and both the web interface and the IPP printer emulation get DNS-SD-advertised, but the printer emulation is not actually working yet.
  • Avahi: Updated the README files of ippusbxd and ipp-usb telling that one needs Avahi 0.8.0 to get the service on localhost advertised or one can alternatively patch older versions.
  • Google Summer of Code 2020: Introduced the students to their projects and helped them to create their proposals. 5 students have done their assignments successfully and posted good proposals so that we will ask for slots for them. There are some more candidates for OpenPrinting who did not convince us and so we will drop them. With these 5 students and the great work of Michael Sweet and Alexander Pevzner we will get most of the work needed for IPP-based printing and scanning done. Tomorrow (March 31) is deadline for student applications at Google.
  • Bugs.


  • Prepared new upload to address a couple of Shell bugs before freeze [diff]



  • Various discussions with OEM to help with driver writing
  • Proposed a branch to libfprint to expose the finger status to the API (and give to drivers the ability of doing it) [MR]
  • Nice win for libfprint-TOD as we got goodix to propose a driver upstream for a future chip.
  • Cleanups to my huge fprintd test suite [MR, merged]
  • Some improvements to the MR to split python tests [MR, merged], reported a meson issue.
  • Updated various fprintd merge requests to address comments [!42, !34, !45, !48, !57, !58]
  • Some random maintainance of libfprint [!126, !129, all merged]
  • Reviews




  • Testing 3.11


  • Updated vte

Snap Store:

  • Debugging issues in the 3.36 branch


  • More work on the patches to the firefox snap
    • Cleanup which reduces the size by 72M
    • Updated to gnome-3-34 snapcraft extension
    • Dropped mimeapps.list to take advantage of portals

snapd dbus activation:

  • I believe snapd PR #5822 (support for user session daemons) is now in a state ready to land. All review feedback from Samuele has been addressed, and I tracked down the intermittent failure in the spread tests.
  • The intermittent failure was caused by another test not properly cleaning up after itself. If it ran before any of my tests, we’d hit a failure. If it ran afterwards, everything passed. I split the fix for this failure out into snapd PR #8385 so it can be landed independently of the user daemons changes.
  • I’m currently working to rebase the top level dbus-activation branch now.

snapd xdg-desktop-portal integration:

  • In the discussion on x-d-p PR #443 it was suggested that it would be useful to have a “file access” command to identify host system paths that a snap can access without being proxied by the document portal.
  • I put together snapd PR #8356 as a start on this. It implements a snap routine file-access command that behaves similar to the flatpak info --file-access command. It has unit tests, but needs a spread test to verify that it works in practice.
  • Once matching support is added to xdg-document-portal, this would mean that e.g. using a portal file chooser to select a file in the user’s home directory would return the real path if the snap had plugged the home interface.

Misc snapd work:

  • The snapd devs have been moving some of the CI workflows over to Github Actions. I made a couple of small contributions in snapd PR #8374 and snapd PR #8375.

CUPS access control:

  • I was tagged into a discussion about security and authentication in the snapped version of CUPS.
  • The approach suggested by @jdstrand is to have cupsd make call outs to snapd to decide whether to process administrative requests.
  • After reading over that thread, I believe a different approach would be better: implement polkit support into cupsd. This would solve a few extra problems (e.g. allow passwordless administration by members of the host system’s lpadmin group), and would likely be useful to all other Linux distros. This second point is important if we want to convince Apple to take the patches.
  • At present, the thread is waiting on feedback from Jamie. I’m willing to give Till further pointers as needed.