Monday 12th November 2018

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 the Community Hub (this site).
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: Monday 5th November 2018 🎇

3 Likes
  • Updated some GNOME snaps to to build with core18 and gnome-3-28-1804:

    • evince
    • gnome-clocks
    • gnome-font-viewer
    • quadrapassel
  • gnome-3-28-1804

    • Refreshed
    • Migrated source to gitlab, building from a launchpad mirror
  • gtk-common-themes

    • Tested fixes that have accumulated in candidate and published to stable

:computer_mouse: Gnome Shell performance (stutter | latency | CPU):

:bug: Bug management:

  • :pick: Backlogs progress:
    • :chart_with_downwards_trend: Good news (falling backlogs): firefox, xorg, gedit, pulseaudio and bluez.
    • :chart_with_upwards_trend: Bad news (long term growing backlogs): ubiquity and xorg-server.
2 Likes

Installer, worked with jibel on new layered image, by:

  • Finish and clean livecd-rootfs changes for creating layered squashfs in a generic way (but without snap and langpacks support for now)
  • Check if negative langpack layers are achieveable size-wise, and yes, they are! :slight_smile:
  • Look at casper, no change needed \o/
  • Investigate debian-cd and ubuntu-cdimage to build a new image system + needed changes for not hardcoding our multiple layers
  • We got a first image built with Jean-Baptiste (see his screenshot, showing a live system with multiple layers mounted). Still works needed to clean it up.

No rls bug assigned

2 Likes
  • Some more work on systemd --user things, should be more or less stable for now, try the PPA in a VM (cosmic, not bionic).
  • autopkgtest:
    • merge some config changes to support a new big package
    • a few retries and stuff for ongoing transitions to help with archive opening (done now)
    • discuss autopkgtest for stable updates policy with SRU team a bit, they are trying to decide when to mark tests as bad and how much investigation to do beforehand
  • Review and merge some MRs for mutter, gnome-shell, gnome-control-center & upload.
  • Help to review @azzar1’s MP for update-notifier
  • Start looking at nautilus update :cold_sweat:

rls:

  • assigned rls bugs are for this upload, which is verified and awaits release (I also checked the tests triggered by it and proposed a hint, which is what triggered the discussion I referred to above), nothing further for me to do here.
1 Like

Disco work

  • merges: desktop-file-utils, alsa-plugins, alsa-utils, geoclue-2.0, orca
  • syncs: udisks2, libical, libgusb, libssh, gutenprint, xserver-xorg-video-intel, alsa-lib
  • updates (some in Debian): gnome-disk-utility, dconf, libmtp

SRUs, Reviews & Sponsoring

  • desktop-file-utils/font warning fix to cosmic&bionic
  • some fixes to the bolt 0.5 SRU to bionic
  • merged my g-i-s/legal url change to disco and SRUed to cosmic and bionic
  • backported a shotwell segfault fix and SRUed to cosmic
  • SRUed shotwell 0.28.4 to bionic
  • sponsored some nautilus fixes from Marco
  • sponsored update-notifier/livepatch fix from andy (disco & bionic SRU)

Other

  • followups on the rls-bb-tracking bugs triaging (investigated some issues, discussed next steps, etc)
  • some extra work on the desktop plans for the D-cycle
  • looked at blocked SRUs and tried to get those unblocked (basically looked at e.u.c to see if those were regressions and talked to people about the issues)
  • updated the version script to track ‘disco’, unblocked ff-tagged bugs (still need to make the ff tag by serie for next cycle)
1 Like

My week:

  • HPLIP: Investigations whether driverless printing instead of HPLIP printing can be preferred. For network printers yes when scanner identification bug is fixed. USB printers need to continue to use HPLIP as with running ippusbxd the scanner cannot be accessed by HPLIP (we would need driverless IPP scanning to do all through ippusbxd).
  • HPLIP: Synced 3.18.10 from Debian into Disco (archive opened today!).
  • CUP/cups-filters: Continued investigation on duplicate execution of page management options (copies, collate, N-up, …). Discussed problem with Mike Sweet, needs fixes in pdftopdf filter and in CUPS’ PPD generator for driverless printing.
  • OpenPrinting web site: Discussed suggestions for a new logo.
  • Bugs.

Till

This was a short week, I was offline on Thursday and Friday.

firefox24 firefox

  • prepared 63.0.1 build4 update − had to pin the version requirement of cbindgen when generating the source tarball, otherwise it would pick a more recent version that caused build failures
  • synced with foundations to request rustc and cargo updates (firefox trunk now requires rustc 1.30.0 and cargo 0.31.0)
  • filed upstream bug #1505593 after @popey’s report, and posted on the snapcraft forum to enquire about it

chromium22 chromium

  • updated dev to 72.0.3595.2
  • reported an upstream problem with source tarballs generation that’s blocking pending updates of stable, beta and dev branches
  • confirmed with the store team that publishing a core18-based snap is A-okay
  • rebuilt stable snap against core18, published in candidate channel (amd64 only for now)
  • discovered that when doing automated builds on launchpad one can request using snapcraft installed as a snap from a specific channel − using snapcraft 3.0 from the beta channel allowed some progress on the issue with build failures on !amd64
  • thoroughly tested the core18-based snap in the candidate channel on trusty, xenial, bionic, cosmic, disco, Debian testing and Fedora 28 (confirmed that a fedora update fixed an error with xdg-open), and issued a call for testing

snapcraft24 snaps

  • confirmed that sogou input works in gnome app snaps (following up on PR #156)

:bug: rls-bb-tracking bugs

  • bug #1754671 in network-manager: pending an upstream decision to backport the fix to the 1.10 series
  • bug #1795169 in firefox: haven’t had time to investigate yet
  • bug #1799230 in libreoffice: sitting in bionic unapproved queue

Nautilus:

  • Fixed some recent-engine crashes [pkg].
  • Upstream fixes to the recent-manager update branch, refactored to be async (now upstream wants tests too, and I’m the lucky first guy who’s starting this new tradition -_-).

GNOME Shell | Mutter:

  • Fixed notification manager to support uris as image-data hint (as per fdo standard) [MR]
  • Some IRC discussions on double-events on X11 touch events fix [PR]
  • Some more tests and updated my fix for this mutter issue, pushing the PR soon
  • Some tests on fractional scaling on xorg, looking at my mutter test-changes

Various extras:

  • GLib: fixed crash on bookmark move [PR] (merged, and cherry-picked to 2.58 and 2.56)
  • XUbuntuCancel SRU, updates branches, reproposed PRs and added gnome-software to the party. [Evrything linked at the bileto ticket, updated too]
  • Some more comments on xorg/xserver!61 (merged)
  • Autocleanups in GtkRecent [MR, cherry-pick] (both merged)
  • Some fixes for nautilus recent not to trigger threads (potentially fixing various crashes in nautilus) [MR, Packaging]
  • Libwnck: MP review, merged and updated my meson branch, then released 3.31.1.
  • BAMF: released 0.5.4 (seems some people want to use it in some core xfce component)
  • Libnotify:
  • Meson:
    • Fix install: stanza in configure_file [PR]
    • Add ability to run specific project tests [PR]
  • Telegram desktop snap:
    • Build using a different API [PR]. Upstream is now controlling evrything! :fireworks:
3 Likes