Review of the weeks 2018/37 – 39

Dear Tumbleweed users and hackers,

Due to my two-week vacation, in which Max Lin took care of Tumbleweed, I’m including the changes of three weeks in this ‘weekly’ review. There were a total of 11 snapshots released in the time since the weekly review 2018/36 (Snapshots 0910, 0911, 0914, 0915, 0916, 0917, 0919, 0920, 0924, 0925 and 0926).

The changes of those snapshots included:

  • Mozilla Thunderbird 60.0
  • Linux kernel 4.18.6, 4.18.7 & 4.18.8
  • Poppler 0.68.0
  • WebKit2 2.22.0
  • Mesa 18.1.7
  • Flatpak 1.0.0
  • LibreOffice 6.1.1 & 6.1.2
  • QEmu 3.0.0
  • KDE Applications 18.08.1
  • KDE Plasma 5.13.5
  • gdb 8.2
  • SQLite 3.25.0

The following things are currently in staging or are building in openSUSE:Factory, leading to the next snapshots to come:

  • Linux kernel 4.18.9
  • KDE Frameworks 5.50.0
  • Qt 5.11.2
  • Mozilla Firefox 62.0
  • GNOME 3.30 (likely not before 3.30.1 though)
  • glibc 2.28; the split of libxcrypt causes some build failures
  • OpenSSL 1.1.1
  • python 3.7
  • ncurses 6.1-20180707: breaks Xen (deprecations added, Xen builds with -Werror)
  • KDE Plasma 5.14.0 (currently, the beta 5.13.90 is being staged for integration testing. WE will only ship once 5.14.0 is released)

Quite a bit of change piled up here. I’m looking forward to keeping delivering you a stable openSUSE Tumbleweed experience.

5 responses to “Review of the weeks 2018/37 – 39”

  1. deliv Avatar
    deliv

    KDE Plasma 5.15.5!!!!!

  2. Dominique Leuenberger Avatar

    Thanks for your notice – this should have been KDE Plamsa 5.13.5 of course. The articel has been corrected

  3. tom Avatar
    tom

    Dominique, thanks for the update! Right now Fedora is my main distro but I’ve been using Tumbleweed in a VM for several months and I’m so impressed that I’m seriously considering to make the move.

    However, one aspects makes me wonder: Tumbleweed is usually very fast in updating to the newest packages. But as soon as you’re having a two-week vacation, it felt like an eternity until kernel 4.18.9 (which fixed 2 vulnerabilities) was offered. In Fedora I got it 10 or 11 days ago, and 4.18.10 (which fixed another 3 vulnerabilities) 2 days ago.

    Perhaps this was an isolated incident – but it makes me wonder if the Tumbleweed project is too much dependent on one person, namely you. I can’t bear thinking about what happens if you decide to take a longer hiatus …

    Can you somehow assuage my sorrow?

  4. Dominique Leuenberger Avatar

    Thank you for your comment. Of course we would be delighted if you were to choose openSUSE Tumbleweed or Leap as your primary operating system.

    As for the speed of submitting kernels, I had a look at the timelines of 4.18.9 and also checked what happened to 4.18.10. Albeit that both are/were not blazing fast, the timelines are also not out of the ordinary, while taking into consideration how much scrutiny a new kernel gets in openSUSE. Kernel 4.18.10 though won’t ‘ever’ reach you as a user: it was submitted by the maintainers on Friday, Sept 28, and the replaced by 4.18.11 on Sept 29 (Saturday). The kernel is currently in a staging area where the entire distro is being rebuilt on top of it (yes, the kernel folks do claim userland compatibility, and very often they are right. ‘very often’ is not good enough for us, which is why we rebuild the entire distro to validate that claim, sometimes having to circle back or being able to report issues. So far, though, none reported 4.18.11)

    If the CVEs the specific kernel versions fixed are deemed highly critical, the SUSE Security team usually reaches out to the Tumbleweed release manager (so, usually me, unless I’m absent) in order to further speed up things as much, possibly skipping some of the pre-integration testings. No such thing did happen though for the last few kernel releases).

    So, even though those kernel versions did not come by ‘very quickly’, taking into account the full stack testing, the delay seems less on the human side than the technical side of things. We might have to verify OBS’ performance again. Its success might make for some performance issues, which would need to be countered with more worker power.

  5. tom Avatar
    tom

    Your answer is highly appreciated! Thanks for your great work. I will test Tumbleweed for another couple of weeks – but I’m already rather sure that I will decide in favor of it.