[prev in list] [next in list] [prev in thread] [next in thread] 

List:       freedesktop-xorg
Subject:    Re: Input region support in the Shape extension
From:       Keith Packard <keithp () keithp ! com>
Date:       2005-01-28 20:05:49
Message-ID: E1CucNJ-0007tZ-Tq () evo ! keithp ! com
[Download RAW message or body]


Around 14 o'clock on Jan 28, Owen Taylor wrote:

> What about the case where the point *is* in the window due to a hole in
> the input region of some overlapping window? I don't quite understand
> why the code in WarpPointer isn't just checking the ancestry of source
> for sprite.win since the x/y being checked is just the pointer position.

Hmm.  A fine question.  Sure looks like we could delete a lot of code this 
way.  It could be that it the sprite trace was added purely as an 
optimization for XYToWindow after this code was designed.

I think I'll have to create some test cases to validate a replacement for 
the WarpPointer checks.  Yay for corner cases.

> (Whether anybody would notice in the remaining lifetime of the universe
> if this was gotten wrong is a different question...)

I've never noticed an application using the source bounds check for 
WarpPointer; surely that's a tiny fraction of the existing tiny fraction 
of WarpPointer uses.

-keith



[Attachment #3 (application/pgp-signature)]

[prev in list] [next in list] [prev in thread] [next in thread]