nanogui: calibration application [& other problems]
Subject:
Re: [nanogui] calibration application [& other problems]
From:
"Massimiliano Turco" ####@####.####
Date:
27 Apr 2004 10:27:03 +0100
Message-Id: <001a01c42c39$dfa1afd0$7e0614ac@tecne126>
Thanks Greg for the support.
I have successfully cross-compiled the nxcal with the libnano-X generated
with my 0.89 version integrated with the 0.90 calibration features.
The problem was I had to include the mou_touchscreen.c and the reference to
IPAQ_MOUSE. And on ARM I use /dev/ts/.
The nxcal binary works on my ARM board and generates a nxcal.dat as
following:
"# Nano-X touchscreen data
# Generated automatically, do not edit
68684 0 -618156 0 67650 -338250 65536
# ---"
Are these values correct?
I am still facing CVS access problems (which don't depend from the online
CVS), so I couldn't yet take a look to the CVS calibration version...
Thanks in advance for your help.
max
----- Original Message -----
From: "Greg Haerr" ####@####.####
To: "Massimiliano Turco" ####@####.####
####@####.####
Sent: Saturday, April 24, 2004 4:22 PM
Subject: Re: [nanogui] calibration application [& other problems]
> : The problem is that on the TFT the cross and the msg ("Microwindows
> : Calibration" - "Touch the cross hairs firmly to"- "calibrate your
handheld")
> : are correctly displayed but if I touch the cross hairs nothing happens
as
> : there is no event.type = GR_EVENT_TYPE_MOUSE_POSITION
>
> You'll need to make a couple more mods to 0.89, check out
> engine/devmouse.c and nanox/srvevent.c looking for the
> MOUSE_RAW status flag. When calibrating, mousedev.flags
> set set to MOUSE_RAW, which tells nano-X to send mouse
> events in raw state to the calibration application. Likely this
> is why you're not getting any events in your nxcal application.
>
> : An other question. I have investigated in the 090 project any reference
to
> : "nxcal.dat" ***without*** result.
>
> Look at the CVS version.
>
> So, in which way the values saved in the
> : .dat file are used by the other application? Those values are
corrections to
> : the mouse position, right? a sort of compensation... any help? any
reference
> : of "how to use" application? I also figure that on PC side there is no
wrong
> : mouse position...
>
> The nxcal application in CVS works by reading touchscreen valuesa
> in raw mode, computing a transform, sending it to nano-X, and
> turning off raw mode. When the server comes up the next time,
> the nxcal.dat transform data is read and the same values used
> in the transform filter.
> :
> : For the problem with cross-compile for ARM the microwindows 0.90, the
error
> : was:
> : In file included from /usr/include/string.h:375,
> : from winmain.c:8:
> : /usr/include/bits/string2.h: In function `__strcpy_small':
> : /usr/include/bits/string2.h:441: warning: pointer of type `void *' used
in
> : arithmetic
> :
> Did you fix this? How?
> :
> : For my problem accessing the Microwin CVS it should be a problem of our
LAN,
> : I hope I will solve it soon...
>
> The CVS is up, please let me know what your access errors are.
> Regards,
> Greg