Preliminary Wayland support

[Darxus]

To work with wayland, gdk_x11_ calls and Xlib calls need to be wrapped in build-time and run-time backend checks: http://developer.gnome.org/gtk3/3.3/ch24s02.html#id1502079

$ grep -r gdk_x11 .
./midori/sokoke.c: Atom save_mode_atom = gdk_x11_get_xatom_by_name_for_display (
./midori/midori-browser.c: gdk_x11_xatom_to_atom (XA_INTEGER),

$ grep -r Xlib .
./midori/main.c: #include <X11/Xlib.h>

http://wayland.freedesktop.org/gtk.html

The wayland backend in GTK+ 3.4 works, so as soon as these are done midori should be usable with wayland.

[gue5t gue5t (gue5t)]

I put some work into getting webkitgtk working with wayland and then tested out midori. We do indeed need runtime checks in a few places. The attached patch is a good start, but we need to do some other work to make Wayland (and other GDK backends) work perfectly with midori, e.g. add runtime checks for the version string instead of always "X11" if it was available at compile time.

To get webkitgtk 1.8.3 to build, I had to reimplement webkit's GtkWidgetBackingStoreX11.cpp in terms of wayland surfaces, and to do /that/ I had to make the symbol gdk_wayland_create_cairo_surface (in GTK3) visible; presumably those two packages will make their own efforts to fix things, but in the meantime this was enough to get midori up and running. It should be noted that libunique3 is not wayland-compatible; I had to pass --disable-unique to configure.

I'll submit additional bug reports for fixes to individual wayland-specific issues, but this is, again, enough to build and run for now.

[Cris Dywan (kalikiana)]

That GDK_IS_X11_DISPLAY looks dubious - why is it hardcoded to TRUE?

[Cris Dywan (kalikiana)]

The #define is there because older GTK+ releases don't have those macros. It's safe to "hard-code" TRUE because it's always inside an X11 check anyway. Wayland here we come!

[Darxus (darxus)]

After talking to gue5t some about this, I don't think it should be marked "fix committed" because the necessary changes haven't been completed or submitted to webkitgtk.

[gue5t gue5t (gue5t)]

I'm not sure what classification exactly this should have, but it's not quite "New". On the midori side our work for preliminary Wayland support is done, but for end-users it's not going to be usable until other projects make some progress (I'm working on getting my webkit patch to work better with the other code so it can be submitted to upstream, but that wouldn't be doable until upstream GDK exposes gdk_wayland_create_cairo_surface). "In Progress" is probably most applicable.