elementary-dev-community team mailing list archive

[Victor Eduardo <victoreduardm@xxxxxxxxx>]

I agree. *pantheon-wallpaper* is still not ready for Luna. All the planned
cool features supporting the effort are yet to be implemented and since
GNOME doesn't depend on Nautilus to render the wallpaper anymore, it would
be a clever move to just use that. Pantheon is just a shell, while GNOME is
a desktop environment and we depend on it anyway.* *Plus eOS will be super
lightweight without it.*
*
I don't know if dropping the project is a good idea, but we definitely
don't want pantheon-wallpaper to be shipped in Luna in its current state.

Also, the most popular applications (Shotwell, etc.) are designed for use
in the GNOME platform, and we want the OS to be compatible with the current
spec. The new desktop-agnostic implementation should be introduced after
Luna.

Regards.

On Thu, Mar 29, 2012 at 8:04 AM, Сергей <shnatsel@xxxxxxxxx> wrote:

> Hello everybody,
>
> it's hard to admit it, but pantheon-wallpaper probably won't make it
> for Luna. It has dreadful bugs like
> https://bugs.launchpad.net/pantheon-wallpaper/+bug/814948 or
> https://bugs.launchpad.net/pantheon-wallpaper/+bug/886633 and no
> active maintainer.
>
> Moreover, I think it's no longer relevant. It was started back when
> Nautilus was drawing wallpaper. Nowadays GNOME wallpaper is drawn by
> gnome-settings-daemon, so the advantage of being standalone is no
> longer relevant. Pantheon-wallpaper uses Cairo for rendering and
> therefore its transitions are smoother, but I've compared them on a
> laptop back from 2005 and I can't say pantheon-wallpaper is *much*
> smoother. In addition, GNOME implementation supports interfaces about
> which we haven't even dreamed about yet, e.g. libvte
> semi-transparency, Ubuntu lock screen, etc, and its way of storing
> configs is already widely supported.
>
> Smooth transitions are a job of GPU drivers. Client-side workarounds
> for that are not a good idea. GDK pixbuf which is [probably] used by
> gnome-settings-daemon is more likely to get performance improvements
> from GPU drivers than Cairo. Latest Intel drivers boost its
> performance many times compared to previous versions, while Cairo
> performance is largely unchanged.
>
> Maintaining a custom solution for the sake of improvement
> opportunities doesn't make much sense either. Most features like
> https://blueprints.launchpad.net/pantheon-wallpaper/+spec/focus-blur
> should be implemented as standalone daemons telling the wallpaper what
> to do, not inside wallpaper service to bloat it with potentially
> unused features.
> And it's very unlikely that we'll drop gnome-settings-daemon in the
> foreseeable future, so the original problem which led to creation of
> pantheon-wallpaper is extremely unlikely to reappear.
>
> Given all of the above, I propose dropping pantheon-wallpaper
> completely. It's not worth the effort.
>
> --
> Sergey "Shnatsel" Davidoff
>
> --
> Mailing list: https://launchpad.net/~elementary-dev-community
> Post to     : elementary-dev-community@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~elementary-dev-community
> More help   : https://help.launchpad.net/ListHelp
>