[PATCH synaptics] support inverted scroll direction

[Peter Hutterer]

On Fri, Apr 13, 2012 at 01:30:42AM -0400, Alyssa Hung wrote:
> On 12 April 2012 23:50, Peter Hutterer <peter.hutterer at who-t.net> wrote:
> > On Wed, Apr 11, 2012 at 10:00:56PM -0400, Alyssa Hung wrote:
> >> This patch adds two options to the synaptics driver: InvertEdgeScroll
> >> and InvertTwoFingerScroll.
> >>
> >> When enabled, they will cause the scroll direction to be inverted when
> >> edge scrolling and two-finger scrolling, respectively. For instance,
> >> moving a finger downward along the right edge will pull the scrolled
> >> content down (causing the page to scroll up). This behaviour is
> >> consistent with scrolling on modern touchscreen devices.
> >>
> >> Options are provided separately for edge scrolling and two-finger
> >> scrolling because they follow different metaphors; edge scrolling has
> >> traditionally been analogous to dragging the scrollbar tab, and may
> >> therefore be sensible to leave uninverted even when two-finger scroll
> >> direction is inverted.
> >>
> >> Signed-off-by: Alyssa Hung <ahung at isisview.org>
> >
> > Thansk Alyssa, two comments regarding this option:
> > In general, I'm not in favour of adding more options to synaptics. This
> > particular one is something that we need to deal with in other drivers as
> > well. With server 1.12, scrolling is now controlled by the increment in the
> > XIScrollClass, simply changing that could change not only the speed but
> > also the direction. While support for modifying the increment is the best
> > solution, it needs a protocol revision.
> >
> > In this particular case, synaptics already has a
> > VertScrollDelta and HorizScrollDelta option that controls the amount of
> > movement before a scroll event is emitted. I think it should be simple
> > enough to change the code that negative deltas are permitted for scroll
> > direction inversion.
> >
> > Cheers,
> >  Peter
> 
> Hi Peter, thanks for your advice.
> 
> I wasn't aware of XIScrollClass increments or their relationship to
> the VertScrollDelta and HorizScrollDelta options. Following your
> recommendation, I have attached a patch that allows them to be set to
> negative values.
> 
> I regret that it won't be possible to invert scroll direction for only
> two-finger scrolling or edge scrolling, but I suppose it isn't a
> common use case for them to both be enabled.

That's fine. We didn't support different deltas and using both scroll
methods with different parameters is not a use case I want to support.
thanks for the updated patch. 

38b93b7..1874094  master -> master

Cheers,
  Peter