On 08/20/2010 03:12 AM, Yang, Jianwei wrote:
Not sure which MultiTouch or MultiPoint protocol standard you are aligning with?
I tried to make that clear in the patch description:
------------
This patch optionally provides a mode of multitouch that works with
Qt 4.7's approach of taking input from multiple /dev/input/event
interfaces. It does so by creating one input device for each supported
contact point (2 in this case)
...
Once the new multi-touch protocol is supported up the entire vertical
stack, I expect this would shift to support that mechanism of exposing
multitouch data.
-----------------
I've verified functionality via the qt-demo application 'fingerpaint'
How about its comparison with
linux-2.6-mid-ref/Documentation/input/multi-touch-protocol.txt?
This patch doesn't provide the ABS_MT_* events. I have another patch I'm
working on that does that, but that is independent of this patch, and also
doesn't work with any user space stack, so is of less use at this point.
If you enable the ABS_MT_* events, Qt starts throwing lots of debug data to
the console due to it not understanding the ABS_MT_* event messages.
The current version of Qt 4.7 that supports the multiple input approach also
doesn't perform the user space finger tracking that is required for the
multi-touch-protocol to work well.
How about your test result on latest MeeGo trunk at AAVA SC?
Patches 1-3 work great; I've tested them on multiple Aava hardware revisions.
Patch 4, when you use only one finger, works fine.
Patch 4, when you use multiple fingers (2), is hit or miss. Sometimes the
second touch point comes through fine, other times the second touch point data
from the i2c interface is only valid for certain ranges of the X and Y axis,
resulting in general screwiness when you press with two fingers on most of
the screen, and good tracking within a small region (usually within the
lower-right quadrant of the upper-left quadrant of the screen)
I originally though the screwiness was resolved in DV2, but subsequent
tests are showing that not to consistently be the case.
I plan to do some more testing on that next week and file some bugs on
http://bugs.meego.com about it. From what I can tell so far, the driver
is doing the right thing, but the hardware/firmware/touch controller is
going bonkers.
James
_______________________________________________
MeeGo-dev mailing list
MeeGo-dev@meego.com
http://lists.meego.com/listinfo/meego-dev
