Re: Porting GNOME to Wayland

[Tristan Van Berkom <tvb gnome org>]

On Sat, Mar 16, 2013 at 3:32 AM, Matthias Clasen
<matthias clasen gmail com> wrote:
> Spring is in the air - things change, people are looking for things to
> try and new goals. I propose that we set ourselves a new goal: port
> GNOME to Wayland
>
> Wayland has reached the 1.0 milestone recently and it has already had
> some good success in the embedded space. Many of us have silently
> assumed that Wayland is the future display system on Linux, and that
> we will get to using it eventually. But to reach its full potential,
> it needs the push of a full desktop porting project. I think GNOME is
> the right project for this and now is the right time for us to embrace
> Wayland.
>
> I am confident that the Wayland and X communities will be able to help
> us in reaching this goal.
>
> Initially, the main work in this effort will be to give gnome-shell
> the ability to work as a Wayland compositor - something that has
> already been demonstrated at last years Guadec in La Coruna. Then we
> need to port functionality for which we've so far relied on the X
> server, such as display configuration or keyboard accessibility
> features. Lastly, the GTK+ Wayland backend needs some love to reach
> parity with the X backend. We will retain the ability to run X
> applications in a compatibility mode, so there is no need to rush and
> port all the worlds applications to Wayland. Porting of applications
> can happen independently and at its own pace.

I'm asking this out of my own curiosity... what kind of porting work
would be required for an 'application' to be ported to wayland ?

Shouldn't that be transparent for most applications by virtue
of linking against the new default wayland GDK backend ?

i.e. usage of the gtk+-3.0.pc would imply wayland anyway
in the bright future where GTK+ is installed on a wayland
capable system, right ?

Will applications need to update configure.ac to specify
a specific GDK backend ? (for systems which might
offer both GTK+ backends, x11 and wayland ?)

And... basically I suppose we're mostly talking about applications
which explicitly #include <gdk/gdkx.h> that might need any
porting, if any applications do need porting ?

Cheers,
           -Tristan

> As far as a roadmap is concerned, I am fairly optimistic that we can
> have gnome-shell work as a Wayland compositor within 6 months. That
> will allow us to have optional Wayland support in GNOME 3.10, while
> still using X by default. Reaching this milestone by 3.10 will enable
> experimentation with Wayland, and should help us to take the next step
> for GNOME 3.12: a fully converted desktop, with no regressions. If we
> realize during 3.12 development that we won't be able to close all
> feature gaps in time for 3.12, it should be possible to keep
> gnome-shell on X for one more cycle without affecting the rest of the
> desktop too much.
>
> This proposal is mine, but it has been discussed with the release
> team, as well as with gnome-shell, GTK+ and Wayland maintainers.
>
> For more details about Wayland see: http://wayland.freedesktop.org
> For more details about this proposal, see http://live.gnome.org/Wayland
>
>
> Let me know what you think,
>
> Matthias
> _______________________________________________
> desktop-devel-list mailing list
> desktop-devel-list gnome org
> https://mail.gnome.org/mailman/listinfo/desktop-devel-list