Re: Patch for updating gtk over wayland interface

[Benjamin Franzke <benjaminfranzke googlemail com>]

For review its always better to git send-email, and since this is
wayland related it should
be fine to send gdk-wayland patches to the wayland-list, maybe with a
gtk prefix.

To comment on the patch i'll copy the relevant parts for now.

>-  wl_display_sync_callback(display_wayland->wl_display, sync_callback, &done);
>+  callback=wl_display_sync(display_wayland->wl_display);
>+  wl_callback_add_listener(callback, &sync_listener, &done);
>+
>  wl_display_iterate(display_wayland->wl_display, WL_DISPLAY_WRITABLE);
>  while (!done)
>    wl_display_iterate(display_wayland->wl_display, WL_DISPLAY_READABLE);
>+

There is a wl_display_roundtrip that can be used now.

>+/* no visual interface for wayland now */
>+#undef WL_VISUAL_HANDLE

Since wl_visuals are really removed, everything that used or handled wl_visuals
should be replaced with the use of egl configs and/or the shm format tokens.
Just temporarily disabling those code paths is ugly.

>-  visual = display->premultiplied_argb_visual;
>+  visual = WL_SHM_FORMAT_XRGB32;

This should give a warning since WL_SHM_FORMAT_XRGB32 is an integer
not a pointer.
The type of "visual" is still "struct wl_visual *" (where the struct
is declared nowhere).
So assigning the new SHM tokens to the old wl_visual variables should
be avoided,
rather use some sth like:
enum wl_shm_format format = WL_SHM_FORMAT_XRGB32.

2011/11/2 Zhao, Juan J <juan j zhao intel com>:
> Hi there,
>
>          These days I built gtk+ in my platform and updated the interface
> and fixing some problems with the help from wayland-devel. Now the first
> step is done: some simple applications in gtk+/test can work well now.
>
>          So I send this patch to gtk-devel for review and asking for
> suggestions.
>
>          As I know, Bnf in wayland-devel also has interest on this and has
> similar patch. So I also add him and ask for suggestion.
>
>          I’m not sure which patch format gtk-devel likes, so I just send it
> with attachment not by “git send-email”.
>
>
>
>          Of course there are still lots of further works should be done, I
> will go on to fix there.
>
>
>
> -----
>
> *^_^* BRs,
>
> Juan
>
>
>
>