I'm installing Ubuntu kiosks in a Paris museum for a Google project, and
we're trying multitouch gestures for the first time on this project.
However the gestures are not often recognized: swiping across the
screens often just drags the mouse cursor instead. I'm seeing
gestures work intermittently, maybe 1/8 times the motion will be
recognized as a gesture, the other 7/8 it just drags the pointer.
That's almost functional, but not often enough to satisfy the museum
directors here.
I've watched the output of `evtest` and I see plenty of X and Y events
in slots 0 and 1. `geistest` output matches the output of `ginn`, and
`ginn` is recognizing all the same gestures appropriately. So it looks
like the GRAIL layer is the one missing the gestures, if I understand
the architecture correctly.
We're testing Iiyama optical touchscreens with the Quanta controllers.
We have a budget, so I can buy different touchscreens if y'all can
recommend better, but we're running out of time.
I'll attach more info recorded from `evemu-record`.
So my kiosk is running a chromium browser window and a Google Earth window. Today I discovered that gestures are always recognized by `ginn` and other tools when performed over the small chromium window, but they are rarely recognized over Google Earth, causing a click and drag instead. So it almost seems like Earth is "capturing" all the input events when it thinks a click and drag is happening. Perhaps this could be reproduced by others?