?? Fri, 14 Nov 2008 14:40:09 +0700 IL'dar AKHmetgaleev ????????:

Hi! As I wrote to gimp-users list I had problems with gtk+ applications, wacom and xinerama same with
http://bugzilla.gnome.org/show_bug.cgi?id=66813

So I have invested some time to get it fixed. This is patch (see attach) which makes wacom work right in my three screen setup.

Please test it with twinview and other kind of multihead setups. Hope to see changes in upstream.

Greetings.

This problem seems to be well known but still not fixed. http://sourceforge.net/forum/forum.php?thread_id=1916305&forum_id=236872 http://bugzilla.gnome.org/show_bug.cgi?id=66813

So once again. I have true xinerama setup with three heads and with different resolutions (1024x768+0+282, 1680x1050+1024+0, 1024x768+2704+282).

I can't use twinview because of I'm using two video cards.

I'm using $ xsetwacom set stylus Screen_No
as keyboard shortcuts to map tablet to one of the screen. I have no problems with X's pointer and with non GTK applications such as blender and pencil (both with pressure sensitivity).

But in all GTK applications (gimp, inkscape and mypaint) in screen mode cursor have offset. Seems GTK maps my tablet to whole virtual screen (3728x1050). Moreover using tablet in screen mode periodically makes my wacom driver hang. My cursor stops and then I can't kill X server and can't switch to console.

I have to use window mode and keep gimp and inkscape windows fullscreen. I still have offset. But it works anyway.

So why GTK uses this strange way to get pointer's position and doesn't uses core pointer's coordinates as blender or QT?

------------- my system:
gentoo amd64
linuxwacom-0.8.0_p3-r1
xorg-server-1.3.0.0-r6
gtk+-2.12.11
gimp-2.4.7
inkscape-0.46-r5
xorg.conf in attach
-------------

Thank you.