Desktop Team Updates - Monday 25th November 2019

Desktop Team Updates
1

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 Update - Desktop Team Updates - Monday 18th November 2019

2

Highlights:

  • Glimpse - 16% of my time
  • gnome-3-34 extension - 48% of my time
    • Worked on build-environment bug. Still not finished. The TL;DR is that the build-environment section is stored in a list of OrderedDicts and hash() can’t be implemented on dicts so some other method needs to be implemented to properly merge such lists.
  • Help out with forum posts - 2% of my time
    • Replied to a couple of posts but really need to do more in this area in the future.
  • cherrytree - 3% of my time
    • snap’d it but still there are theming issues. need to look more into this.
  • compliance training - 3% of my time
    • did several but there are still 7 left
  • Misc (everything else) - 28% of my time
    • opened my first MIR! but there is still another one needed for libscalar-list-utils-perl
    • looked a bit into harfbuzz/pango in update_excuses. pango still has an error that will need to be reproduced locally and reported upstream
    • filed gnome-shell bug that is hitting me frequently on both desktop and laptop. would like to find time to look into a reproducer.
    • provided feedback on Igor’s KDE extension blog post and new docs page
    • looking into why gjs is added to the extension layouts and causes 500-1000ms latency. Turns out it was added for convenience but nobody knew it would affect the performance like it does.

Plan for W48
It’s a short week for me, working only Monday and Tuesday due to the holiday + swap days.

  • gnome-3-34 extension
    • Fix this darn build-environment bug - I’m soooo close!
  • Compliance training - finish up the rest
  • update_excuses:
    • file MIR on libscalar-list-utils-perl
3 Likes
3

zsys: (with jibel)

  • implement authorization framework on zsys, via unix socket SO_PEERCRED element passing and polkit.
  • hooked it up in GRPC (apparently, no-one has ever done it / code published before \o/)
  • added Version() grpc call, to validate non authorized (permissive) calls. The version is automatically extracted when building the package (and can be separated between daemon and server)
  • adapted the packaging to new daemon/client architecture
  • added backward compatibility for eoan users.
  • uploaded grub2, grub2-signed and grubzfs-testsuite with multiple fixes (see last week), pet along autopkgtests.
  • uploaded new zsys 0.3. However, currently FTBS due to golang-go not providing a valid alternative. Will revert to 1.12 and fix missing system dbus failing tests in builder environment.

Misc:

1 Like
4
  • See @didrock for ZFS
  • Verified images of Focal with latest ZFS and grub uploads
  • Fixed automated upgrade tests.
  • Maintenance of the QA lab.
5

firefox24 firefox

thunderbird24 thunderbird

  • tested and validated thunderbird 68.2.1+build1 and related dependencies in {bionic,eoan}-proposed, now pending publication by the security team
  • updated to 60.9.1 for a fix to bug #1850651 (Authentication constantly re-requested for Google); since this was not a valid candidate for publication by the security team, I prepared a SRU and uploaded to {xenial,disco}-proposed

chromium22 chromium

package24 other

1 Like
6
  • one day off
  • spent some time post travel to catchup on things
  • created trello cards for next cycle roadmap items
  • looked a bit at the librsvg build, tried to backport some upstteam fixes and then handed it to Olivier who knows better the build stack used there
  • Tested gamemode and fixed some packaging issues
  • spent some time try unblocking some packages from proposed (retried some tests, uploaded fixes to openscad/gegl/…)
  • Updated duplicity, fixed some python3.8 deprecation warning, reported the remaining build problems upstream
  • finished the requested Canonical online training
  • wrote to discourse/mailing list about enabling integrity check by default, tested performances on some machines as part of the proposal
  • sponsoring (vala)
1 Like
7

GNOME and friends

  • Merged mutter fix for crash on weird x11 events, and cherry-picked to stable [MR]
  • Deep review of libfprint v2 APIs and code changes:
    • Reworked the way we handle the user data on state machines
    • Ensuring we provide async callbacks and can chain operations
    • Rewritten all the missing device events, and added one for managing device prints
    • Various cleanups
    • Got commit bit on libfprint!
  • Even more libfprint cleanups [MR]
  • Support on state machine delayed actions [MR]
  • Support variadic errors functions [MR]
  • Some reviews
1 Like
8
  • Snap-only CUPS in 20.04: Investigations for whether this is actually doable. Especially checked the documentation of SANE and PWG’s driverless IPP scanning standard to evaluate whether it is possible to code the IPP scanning support for 20.04. This would be needed to snap HPLIP as it is a driver for both printing and scanning.
  • Snap-only CUPS in 20.04: Posted an e-mail with a summary of what will change in printing and scanning in the future and what is needed for 20.04.
  • Google Cloud Print: Google will shut it down by the end of 2020. Should it still got to 20.04 LTS?
  • Bugs.
9
  • Very short week: three days off
  • Followed up on some internal discussions from the Canonical event two weeks ago
  • Tried to get an lxd runner in the staging autopkgtest environment, blocked on some quota issues (worked with IS a bit on that)
  • Some gnome-session upstream reviews / discussions
  • Started working on a potential systemd patch to not kill processes when gnome-shell exits abormally, still kind of at the research stage on this one
  • Some internal training
1 Like
10
  • Fixed up some regressions from snapd-glib Meson migration.
  • Worked on GNOME Control Center improvements that uncovered some memory leaks.
  • Worked on a native Dart D-Bus implementation.
  • Spent some time trying to improve Snap use of AppStream IDs.
  • Looked into GNOME Software master suitability to bring to focal.
11

libreoffice24-margin LibreOffice

GNOME
GNOME1024×1024 14.6 KB
GNOME build/plaform-snap

Unknown Dart

12
  • Worked on various snap build failures due to requiring new libs. Updated to use the build snap and debug issues with using the build snap
13

Pulse Audio Snap policy module backport (bug #1781428):

  • The backports are now in proposed for xenial and bionic.
  • @jdstrand also ran through the test plan from the bug report verifying that these packages function correctly.

Automatic theme installation for snaps:

  • I attended a meeting with some snapd core devs about how we will handle installation of theme data for snaps not covered by the gtk-common-themes snap.
  • The meeting was constructive, and I published my notes from the meeting extended with some thoughts I had the day after:
  • In short, this would require a background user session daemon that monitors theme changes via GtkSettings, and would call into a new snapd API to ensure snaps to support the theme selection are installed.

Other snapd work:

  • Still working on the user session service management code. I need to get some of the PRs merged.