multi-touch-dev team mailing list archive
-
multi-touch-dev team
-
Mailing list archive
-
Message #00752
Re: Single-touch behaviour
On Tue, Feb 22, 2011 at 10:16:35PM -0500, Chase Douglas wrote:
> On 02/22/2011 09:14 PM, Stephen M. Webb wrote:
> > A discussion of the expected behaviour of one-touch gestures came up
> > earlier today in a telephone conversation. The Unity developers
> > expressed concern that would be unable to implement single-touch hover
> > behaviour because single touches are interpreted by a lower layer as
> > mouse pointer actions are are not delivered as gestures even when
> > requested.
>
> I'm having a hard time understanding the issue; I think I need more
> details :). Are we talking about a touchscreen or a trackpad? What do
> you mean by single-touch hover behaviour? We don't have any devices that
> actually support touch hover events.
Yes we do - the pen on a Dell XT2 for instance. A hovering pen sends
ABS_X/ABS_Y events during hovering, which makes the cursor follow the
position of the pen on screen. When the pen is pressed against the
screen, a BTN_TOUCH event is sent together with additional ornaments
such as pen type. When the pen is lifted, the button events are sent
in reverse, and the ABS_X/ABS_Y events continue to be sent until the
pen is far enough from the screen.
> In 10.10 and soon to be 11.04, utouch sits in the server before any
> pointer events are generated. When gestures are found, no pointer events
> are delivered.
>
> When we get a client-side version of the utouch stack working, we will
> still be grabbing touch events before they are converted into pointer
> events for touchscreens. The pointer will actually move with the first
> touch on the screen, but X motion and button events will not be sent if
> a gesture is found. Touchpads are more nuanced, and I don't really want
> to get into it unless it's germane to the discussion.
It seems to me they have a lot to do with each other.
> So, does any of this clear things up, or have I gone on a tangent? :)
>
> > This issue was also raised today by the community in bug #723304
> > <https://bugs.launchpad.net/utouch-geis/+bug/723304>.
>
> That looks to me like a plain bug. How is it an architectural issue?
Trying it out just now with grail-1.0.19 and XI2.1, listening to tap
only, I can drag the cursor, but the initial portion of the drag is
lost, thus starting in the wrong place. When listening to only drags,
and can tap to move the cursor, but it will only happen 70% of the
time.
Henrik
References