Desktop Team Updates - Monday 26th July 2021

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 .

Last week’s notes are here: Desktop Team Updates - Desktop Team Updates - Monday 19th July 2021

1 Like

Ubuntu Desktop Installer






  • Work on fixing daily build and continue writing build script.


  • Internal event, attending some sessions
  • Client work debugging
  • Attended many sprint sessions, took up most of the week
  • Got the ubuntu KPIs moved to a new cloud
  • Did a bit more integration of unavailable but visible ESM updates into update-manager. Trying to grey them out now.
  • CUPS Snap: Waiting for @ijohnson’s pull request on snapd for the cups printing interface to get merged by the snapd team.
  • cups-filters: Let the PPD-caching function of libppd auto-create an array of presets (sets of option settings, like “PrintQuality=High PrintDirection=UniDirectional ColorModel=sRGB”, which get applied on a given event) from any PPD to be applied on combinations of IPP settings of grayscale/color, print quality, and content type (text, photo, graphics, …). This is for the CUPS printer driver retro-fitting framework so that all the old printers can be sufficiently well (allowing quick, toner-saving draft and also high quality photos, mono/color switching, …) with the simple, standardized IPP interface instead of with the complex, printer-specific PPD interface. This allows easy printing on all these old non-driverless printers even from a phone!
  • PAPPL: The problem in PAPPL mentioned last week turned out to be caused by my PPD-driver-retro-fitting code overloading a fixed-length array with too many items (PAPPL allows only 4(!) choices for the printing resolution). Fixed by adding an appropriate auto-selection method for which resolutions to import from the PPD.
  • Retro-fitting of classic CUPS drivers into Printer Applications: Continued on the driver retro-fit library code. Trying to retro-fit Gutenprint PPD files I ran into the very strict limitation of PAPPL only allowing 4 choices for the printing resolution. Created an algorithm to select the 4 most suitable resolutions from the PPD file. Next challenge was to make the IPP attributes print-color-mode (gray/color), print-quality (draft/normal/high), and print-content-optimize (auto/photo/text/graphics/text&graphics) work with as many of the > 10000 existing PPDs as possible, which requires an automatic algorithm to associate the correct option settings of the PPDs with these attributes (see cups-filters above). Also auto-select the print-content-optimize attribute using the metadata of PDF input files telling with which app the PDF got created. Now one can control 99% of the printers well via simple, standardized IPP options (especially also from phones) and (Update) retro-fitting of all (~10000) PPD files which are usually available in Ubuntu works!!
  • Google Summer of Code 2021: Continued mentoring the students.
  • Linux Plumbers Conference 2021: Proof-read the announcement and the descriptions of our OpenPrinting micro-conference. Both are published now: Announcement, Description
  • Bugs.
  • Spent most of the week in roadmap meetings
  • Sponsored some fixes from oem
  • Trivial packages updates
1 Like
  • LO 7.2.0 snap:
    • continued debugging (current source, current build failure). debugging included making a patch for skia (external dep grabbed during the LO build) and required a lot of tweaking (to get past the skia build failure), and now have a firebird (another external dep LO grabs) failure but not stuck.
    • also a parallel investigation with local builds on the necessity of reverting the upstream commit that causes the build to fail. Friday I made a breakthrough and the commit doesn’t need to be reverted but rather changed and I have ideas on how to change it… Note that the 7.2 builds on LP are still reverting this upstream commit just to get past the build failure for now.
  • five-or-more: finished, tested, and merged. @kenvandine, thanks for merging! I borked the lxd container that the GH action/runner uses and no idea how to clean it remotely - so if any team members know how to reset that container, please let me know :slight_smile: actually a new pr with new changes spawns a new container so mystery solved … Also when writing this, i found and fixed some missing changes.
  • Attended most of GUADEC wednesday and friday, also hosted a Q&A session with @3v1n0 (thanks so so much!) - the video of that session is the first hour of day 3 footage



  • Uploaded i386 (and devhelp) fixed pango to debian
  • Improved gtk4 packaging, adding media modules, libcloudproviders (debian only) and sysprof support (uploaded to NEW by @laney).
1 Like

snapd theme support:

  • Worked on a pull request to expose snapd’s theme installation API to snaps without plugging snapd-control. The code is currently in draft state as snapd PR #10571, needing some extra spread test coverage. The existing test made use of an existing test snap published to the store that doesn’t plug the new marker interface.
  • There were further discussions about improving theming of snapped applications on various desktops last week. As part of that, I’ve been working to improve the appearance of GTK based snaps on KDE based desktops. This requires both snapd and Snapcraft extension changes, so will only gradually roll out to snaps as they are rebuilt (assuming they use the extension).

Ambiance/Radiance theme issue:

  • We discovered some performance problems with popup menu animation with snapped apps using the new core20 based GNOME platform snap and the old Ubuntu Ambiance/Radiance themes. I’ve identified a particular property that removes the popup slowdown when commented out. It’s not quite clear why that makes a difference, as the property is overriden for all nodes in the popup.