[PATCH xf86-input-synaptics 1/8] Add another third state to TouchpadOff for disabling all but button clicks

[Peter Hutterer]

On Tue, Feb 25, 2014 at 02:25:50PM -0800, Carl Worth wrote:
> Hans de Goede <hdegoede at redhat.com> writes:
> >  0	Touchpad is enabled
> >  1	Touchpad is switched off
> >  2	Only tapping and scrolling is switched off
> > +3	Only cursor movement is switched off

whoopsi, i've pushed a patch to fix up this part of the documentation

> I like the idea of this one, and have found it extremely handy already
> trying to get my new Lenovo laptop (sadly, without dedicated trackpoint
> buttons), somewhat usable.
> 
> But it's not clear to me from the above documentation whether things
> like two-finger scrolling are disabled in mode 3. It would be handy for
> me, personally, if not. That is, I would prefer to have no cursor
> movement, ever, but it would be handy to still have two-finger
> scrolling.
> 
> In my testing of this patch, it seems that the scrolling is disabled as
> well, and that's unfortunate.
> 
> It's not clear to me if its appropriate to override this single
> "TouchpadOff" variable for all of these modes. Might it be easier to
> satisfy users with independent disables for each thing? (Dont' we even
> already have enables for things like tapping and scrolling?)

yes, but they are for configuring the feature. the touchpad off property is
to temporarily disable the device without actually disabling the device as a
whole. Think of it as the main power switch, while the rest are dials that
are left in the same state once set.

the "off" property is mainly used by syndaemon for the disable-while-typing
functionality.

state 2 was introduced in 2005, presumably to avoid erroneous events again,
while allowing movement of the cursor.

> So maybe all that's missing is a disable that affects cursor movement
> and nothing else?

the first iteration of this patch did just that, but we found that we'd get
erroneous scroll events. Remember that synaptics supports two-finger
scrolling as well as edge scrolling, it's a lot easier to get invalid events
from the latter.

Also, he current plan is still to switch the touchpad back into normal mode
based on a set of triggers (that we haven't fully identified yet). The
touchpad should work at the same time as the trackstick, similar to the
disable-keyboard-while-typing approach.
So the real goal here is to detect valid scroll events, switch the touchpad
back into normal mode and process them as normal. You'll need something that
switches the touchpad back once you use the trackstick, but that's being
worked on too.

Cheers,
   Peter