Idea: adding an input mask extension

[Keith Packard]

Around 18 o'clock on Dec 26, Anton Markov wrote:

> What if we use a similar idea, but instead of defining the area to be 
> rendered, the new bitmap would define the area of the window that 
> accepts input. Areas which are not set to "on" will never receive input 
> (more specifically mouse events). Instead, the event will go to the 
> window below it.

I've wanted to do this; it looks like a relatively simple change to the 
DIX layer will accomplish this effect.  My thought was that we'd bump the 
shape extension version and have the new version offer an 'input region' 
in parallel with the border and clip regions currently provided.

If you look in events.c, you'll see only a very few places where the shape 
of the window affects input.  Right now, those all use the borderClip 
region.  Making these (conditionally) use a new region would let us 
program that region easily enough.  I think the changes to implement this 
new input region are straightforward.

If anyone is interested in exploring an implementation, I'd love to kibitz.

-keith


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.x.org/archives/xorg/attachments/20041226/13eaf30b/attachment.pgp>