Hello,

I'm trying to implement RANDR support for the Tiny-X i810 driver, at
least so it can display depth 16 and 24 windows at the same time.

Thus I create two layers, one of kind LAYER_FB and one of kind
LAYER_SHADOW. Currently, my framebuffer is of depth 16, and the shadow
buffer is of depth 24; I present an additional visual of depth 24 by
hacking in an extra call to fbSetVisualTypesAndMasks.

It works pretty ok, except for clipping problems between windows in
the different layers. Sometimes drawing operations to the depth 24
window are not clipped if a depth 16 window partially covers their
window.

In my scenario I have a depth 24 window that is updated on expose
events using a CopyArea function, and I'm moving around a depth 16
(the default) xterm window to partially cover it.

If the xterm completely covers the test window in the Y direction, but
partially covers it in the X direction, clipping works as expected. As
soon as I move the xterm vertically so that it only partially overlaps
in the Y direction, drawing operations don't seem to be clipped at
all.

Is there a bug in the layer code, or did I forget to do something? I'm
building from the cvs tag xf-4_1_99_1, but I 'backported' the copyclip
argument transpose error in layergc.c from the latest CVS and can see
no other significant differences, at least in the layer code.

Regards,

Pontus

-- 
Pontus Lidman               Email: [EMAIL PROTECTED]
Nokia Home Communications   Phone: +46 (0)13  4611201
Diskettgatan 11             Fax:   +46 (0)13  4611001
583 35 Link�ping, Sweden    
_______________________________________________
Xpert mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/xpert

Reply via email to