[PATCH libinput 0/5] tablet: change to an axis-specific API

[Hans de Goede]

Hi,

On 14-12-15 03:00, Peter Hutterer wrote:
>
> This patchset changes the Tablet API from an API where we have a few
> functions that take an axis type into one where we have an API that is
> specific to each axis type.
>
> The one-size-fits-all approach turned out to be too inflexible to be useful
> and forced a couple of special API behaviours, e.g. get_value for a relative
> wheel must always returns 0, has_axis for x/y is always true, etc.
> Changing the API to be per-axis is a lot more expressive in code and allows
> us to document and adjust the specific behaviour for each axis accordingly.
>
> Rather than libinput_tablet_tool_has_axis(tool,
> LIBINPUT_TABLET_TOOL_AXIS_PRESSURE), the new call is now
> libinput_tablet_tool_has_pressure(tool).
> and a similar change for the events.
>
> The functionality largely stays the same, but there are a few notiable
> differences:
> * has_x() and has_y() is not available - it wouldn't be a tablet if it
>    doesn't have those
> * has_tilt() instead of has_tilt_x() + has_tilt_y(). We don't init devices
>    with only one tilt axis as tablet.
> * delta support (including delta_discrete) was dropped for all but the rel wheel.
>    delta_discrete was already 0 for all but the wheel, the others were simple
>    delta = old - current. We can punt that to the caller.
>
> The last change enables us to introduce a get_dx/get_dy in the future for
> accelerated motion data - required in relative mode.
>
> Note that this is just the change to the public API. The internals work the
> same as before, but will be adjusted once we agree on the public API.

I've quickly looked over this set and I cannot find anything wrong with it:

Acked-by: Hans de Goede <hdegoede at redhat.com>

Regards,

Hans