[PATCH/neod] make neod more relexable about keycode returned from the touchscreen

Neil Brown neilb at suse.de
Tue Sep 2 02:16:31 CEST 2008

On Monday September 1, mickey at openmoko.org wrote:
> Am Freitag, 22. August 2008 09:45:50 schrieb Neil Brown:
> > Be more relax in keycode used for detecting activity on touch screen.
> >
> > Different drivers may report different key code.  We don't really
> > need to care which keycode is used, just if there was a key press
> > at all.  So use a range test to look for any 'tool' like button press.
> >
> > Signed-off-by: NeilBrown <neilb at suse.de>
> applied, thanks.


In the interests of full disclosure, I'm no longer certain that I will
be pushing the BTN_FINGER -> BTN_STYLUS change.  But the future is an
uncertain thing and having this patch in place is probably a good idea
anyway and will be invaluable if we do go that route.

I am told by the designer of the input layer that BTN_STYLUS does not
mean "The stylus is on the touchpad" as you might expect, but rather
"the button on the stylus has been pushed".  So changing to
BTN_STYLUS, while being very practical, is not really correct.

The alternatives are:
 - leave the patch in mousedev.c which explicitly excludes "ABS_X +
   ABX_Y + BTN_TOUCH" from being treated like a mounse.  I'm not keen
   on this.
 - patch the X server to not open /dev/input/mice at all.
   This would mean that you cannot e.g. configure a bluetooth mouse
   and use it.  However as the Xserver is run with -hide-cursor 
   an external mouse would be pretty useless (I know, I've tried!)

None of these solutions are perfect, but we are getting closer.


More information about the devel mailing list