Re: [Linuxwacom-discuss] wacom_drv: Reemit events from CoreKeyboard

[Cedric S. <ma...@gm...>]

On Thu, Jan 06, 2011 at 03:09:31PM +1000, Peter Hutterer wrote:
> On Mon, Jan 03, 2011 at 12:10:01PM +0100, Cedric Sodhi wrote:
> > I've also tried reattaching the Pad-Device to the XInput Core Keyboard:
> > 
> > $ xinput float 13
> > $ xinput reattach 13 3
> > X Error of failed request:  XI_BadDevice (invalid Device parameter)
> >  Major opcode of failed request:  141 (XInputExtension)
> >  Minor opcode of failed request:  43 ()
> >  Device id in failed request: 0x17
> >  Serial number of failed request:  17
> >  Current serial number in output stream:  18
> > $
> > 
> > I think that would be the most elegant solution. Any chance that
> > reattaching as a keyboard could be made possible?
> 
> the server is configured that key events from a device attached to a master
> pointer (usually the virtual core pointer), will be sent through the paired
> master keyboard (i.e. virtual core keyboard in the default case). so the
> above error isn't really a problem, key events will be routed correcly.
> 
> Cheers,
>   Peter

Hi Peter,

could you alter the code so that the events are sent throught the Master
Keyboard EVEN IF the Pad-Device which runs as a pointer is FLOATING?

The problem is that as soon as I touch the pad, it emits motion data for
pointer device which cause the application (GIMP) to switch the current
device. I can only prevent that from happening if the pad does not emit
motion data as a Pointer device, which, in turn, is only possible if the
device is floating, does not emit any data as a pointer, is not a
pointer at all.

The last option is generally not feasible because the pad must be able
to emit motion data, the second is equally uneligible and the first
seems reasonable, but when I float the pointer it doesnt emit keyboard
events any longer.

Do you see the dillemma? Hopefully you can help.