Patch to mouse accel code

Jim Gettys Jim.Gettys at hp.com
Thu Nov 11 11:47:55 PST 2004


It would be better if the algorithm was coded to avoid floating point,
if it isn't too hard.

We do run on systems without FP hardware (e.g. ARM) based systems, and
minimizing FP is goodness.  This is not an absolute requirement; scratch
your head for a bit and see if it is not too complicated (you aren't
doing many operations, so I don't think it is a big hit if it remained
floating point) to do in integer.
				Thanks,
					- Jim

On Thu, 2004-11-11 at 20:38 +0100, Jan Brunner wrote:
> I wrote a patch that uses the same method to prevent accumulation of 
> rounding errors in the threshold acceleration code as the exponential 
> acceleration does.
> 
> The main advantage of this method (compared to not saving the 
> remainders) is that slowing down the mouse, e.g. "xset m 1/3 1" works 
> without problems (with the current implementation, the pointer doesn't 
> move when the mouse is moved slowly). As mice get higher and higher 
> resolution, this will be appreciated by many users.
> 
> I created a Bugzilla entry and attached the proposed patch. It would be 
> nice if someone with commit access could review it and add it to cvs.
> 
> http://freedesktop.org/bugzilla/show_bug.cgi?id=1688
> 




More information about the xorg mailing list