acalib/Input woes

We have a working prototype for a product that uses a touchscreen.

Due to packaging problems, we need to flip the touchscreen connector
around to prevent it from being twisted in the final product.

When flipping the cable, the origin (0,0) of the touchscreen is
relocated from the upper left corner to the lower right corner
of the display.

I thought that acalib and Input would be able to handle this
situation, but perhaps the don’t.

When flipping the connector, calib won’t run to completion. The
behavior is, all the targets can be touched, and the absf file
is written when the final center target is touched, but when the
“press to complete calibration” button comes up, touching it
does not cause calib to exit - instead the timer runs out and
calibration restarts.

Running calib in verbose mode shows that the coordinates for
the completion button press are always mapped to 0,239. Dumping
the raw coordinates shows that they seem to make sense - the
final button press is certainly within the touch region defined
by the four corners.

Shouldn’t this just work? Any ideas of what we can look for?

Ken Schumm <kwschumm@qsolv.com> wrote:

We have a working prototype for a product that uses a touchscreen.

Due to packaging problems, we need to flip the touchscreen connector
around to prevent it from being twisted in the final product.

When flipping the cable, the origin (0,0) of the touchscreen is
relocated from the upper left corner to the lower right corner
of the display.

I thought that acalib and Input would be able to handle this
situation, but perhaps the don’t.

When flipping the connector, calib won’t run to completion. The
behavior is, all the targets can be touched, and the absf file
is written when the final center target is touched, but when the
“press to complete calibration” button comes up, touching it
does not cause calib to exit - instead the timer runs out and
calibration restarts.

Running calib in verbose mode shows that the coordinates for
the completion button press are always mapped to 0,239. Dumping
the raw coordinates shows that they seem to make sense - the
final button press is certainly within the touch region defined
by the four corners.

Shouldn’t this just work? Any ideas of what we can look for?

Can you contact heather@qnx.com in tech support. She’s been dealing
with some similar touchscreen issues recently, and can do most of the
preliminary work to figure out what’s happening.

She’ll want to know what version of calib and Input you have, and
will probably give you a matched set to try.

Apparently, the calib/touchscreen stuff changed behaviour at some
point and if you mismatch the calib utility and the driver, things
don’t work correctly.

Previously, pete@qnx.com wrote in qdn.public.qnx4:

Ken Schumm <> kwschumm@qsolv.com> > wrote:
[…]

Shouldn’t this just work? Any ideas of what we can look for?


Can you contact > heather@qnx.com > in tech support. She’s been dealing
with some similar touchscreen issues recently, and can do most of the
preliminary work to figure out what’s happening.

Will do, thanks.

[…]