multi-touch-dev team mailing list archive

[Henrik Rydberg <rydberg@xxxxxxxxxxx>]

On 09/18/2010 05:26 AM, Henrik Sandklef wrote:
[...]

> thanks ... hope to see you at FSCONS (fscons.org) in November. Carlos G is
> coming over to talk about M-t.


Thanks for the invitation.

>> Har du kollat in uinput?

> 
> I remember looking at it (after having checked the source code) some years ago.
> Before I spend time on the wrong things, I'd like to check with you if uinput is
> easier/better to use. Generating pointer[1] input events using swinput is done
> like:
>    echo "u 1" > /dev/swmouse4
> which will cause the 5:th 'mouse' to move up 1 step. So it's fairly straight
> forward to use.


The uinput can be used to create arbitrary input devices, but it would not help
you create special devices like the ones you describe.


> I wrote swinput[2] (to test Xnee) some years ago and haven't really looked into
> any other similar stuff since then.
> 
> Do you think that uinput is a better way to go than swinput?


I asked because we are making good use of uinput to emulate kernel interfaces,
and I was just about to launch this project:

   http://bitmath.org/code/evemu/

The goal here is to be able to debug multitouch device problems remotely, as
well as provide a framework for gesture timing measurements.

If the main purpose of your library is to emulate X events, there may be no
particular point worrying about uinput.

Cheers,
Henrik