Sunday, March 25, 2012

[GDM 3.2.1] May I stop maintaining gdm ? :(


Today GDM 3.2.1 seems too hard to maintain on personal PPA anymore... :(


Yeah in fact Fedora17 would be released "completely without any ConsoleKit" but providing "Systemd" as new features... So now git branch goes along with this change dramatically from a couple of month ago. In opposite, Ubuntu still needs ConsoleKit even if it has some security disadvantages... Additionally, Gnome-Keyring has been updated to 3.3.92 as well.

Exactly today my gdm-3.2.1 have broken on Precise Pangolin with latest gnome-shell codes. On meantime, I've tried "tons of" patchworks while keeping ck for gdm-login routines, but still no luck... :(

I won't make any security-holes by dirty hacks, and now lightdm seems well on multi-display configurations, on user-switching, and so, even though we might have to make some new greeter to fit to gnome-shell's appearance and/or Look&Feel... I wonder it might be even easier than keeping gdm3 alive in the future of Ubuntu...

How it should be...?


  1. Whilst giving up on gdm is a pity after all your hard work, I think that your decision is the correct one.You should apply your time to projects that have a chance of working in the future.

    The situation with Ubuntu and Gnome can only get worse as time goes on. When Gtk and Clutter are merged (already half implemented), it is doubtful whether Unity will be able to use Gnome as a base OS at all. It will either fork at 3.2/3.4 or move rapidly to Qt. So losing Gdm is just the first stage of losing Gnome on Ubuntu.

  2. Hi duncan. :)

    Yeah that's right unfortunately...
    As time goes by, basic Gnome components seem to work difficulty without patches for debian/ubuntu. So I suppose to be better to create new projects of login-greeter to be fitted to Gnome3 appearances what runs as lightdm greeter. Yep, lightdm has much more "simplicity" than GDM exactly. :) And now works pretty good on our beloved Precise Pangolin and even in the future as well I hope...

    Finally Qt4 seems fantastic!
    Rich interfaces, Faster renderings, C++ based codings, and so much... I love it, too.
    I wonder why Ubuntu devs didn't take Unity-2D up to "default" dekstop... And I wish Qt4 could be coded by Vala as well in more easier ways... Because Vala could improve great features as "Performance" enhancements you know...

    Anyway we strongly need many types of desktop experiences which we could "choose" in many ways for the purpose of our happy linux life, right? :P


    1. Whilst I agree with your argument that we need a variety of desktops, Canonical seems to be aiming to become a single Unity-only distribution. Xubuntu is leaving (rebasing on Debian for 12.10), Kubuntu has had its official support removed and will probably follow. Finally, the Ubuntu base is now so heavy, that attempts at a light-weight product like Lububtu are becoming almost impossible. By 13.04 Ubuntu will be effectively a Ubity-only distribution, and lightdm will be what it was always intended to be Unity-dm.

      The central issue with gdm is that is expects a systemd environment, its fallback to consolekit is flaky because its an after-thought. Canonical has heavily funded the development of lightdm exactly so they can stay with upstart/consolekit etc. The disparity between a Unity and Gnome environment will only continue to grow.

      With the improvements in Unity 2D, I suspect that Unity 3D will disappear, and you will be left with the Canonical boot-strategy, Canonical dm, and a Canonical desktop based on Qt. It is the natural progression of their current strategy and has the advantage focusing their limited programming resources all in one place. As for Vala, you will have to do the work for yourself to get it and Qt to work well together. Vala was specifically designed to convert from a high-level language to efficient Gobject-C. Its raison-detre is to support Gnome/GObject/Gtk.

      So, why did they not make Unity-2D the default? Because they were scared of the reaction. The introduction of Unity lost them 15% of their user-base, the loss of Kde/Xfce will lose them another 15-20%. Of those remaining, almost half are Unity-hating Gnome users, whilst the other half use Unity either because they like it or simply because it is the default. Changing to rapidly would have increased the likely-hood of this last group moving to another distribution, before CanonicalÅ› solution was ready. They need time to develop a full Qt replacement to all the Gnome facilities which they use to support Unity.

      In short, the variety of linux desktops will continue, but not at Canonical.