nanogui: Thread: Touchscreen pointer disappearing when moved first time

After a long hard battle I finally got the nano-X cursor following the 
pointing device. I had to change the x and y coordinates in 
driver/mou_ucb1x00.c, 0.89, by adding the following code to the 
PD_Read() function.

        temp = event.x;
        event.x = event.y;
        event.y = 1000 - temp;

This now got the cirsor actually following the input device instead of 
being 90 degres out of "phase".

Cool say I and I move back to the cvs download to achieve the same 
thing. But, trouble right off of the bat. When starting cvs nano-X the 
cursor, ie arrow, did not even appear! I checked to make sure I had 
TUXSCREENMOUSE = Y in config and then after several minutes of fussing 
and fuming I find that
GdHideCursor(&scrdev);
had been added to the PD_Open() function in drivers/mou_touchscreen.c. 
Why, I have no idea. So I removed it and added my "fix" code plus a 
printf debug to once again make sure my adjusted coordionates are right 
and rebuild.

Start the new nano-X and BINGO, I have the cursor. Then my joy is dashed 
when I touch the screen to move it and check out my fix. It disappears 
again. I suspect there isd another added GdHideCursor() someplace. If 
that is it, where is it? My debug prints event coordinates that keep the 
cursor in teh middle of the screen, but no cursor.

tj

: PD_Read() function.
: 
:         temp = event.x;
:         event.x = event.y;
:         event.y = 1000 - temp;
: 
: This now got the cirsor actually following the input device instead of 
: being 90 degres out of "phase".

Is this code still required in CVS version?  You are running
in portrait mode, right?  The CVS code automatically swaps
x/y when in portrait L and R modes.

: 
: Cool say I and I move back to the cvs download to achieve the same 
: thing. But, trouble right off of the bat. When starting cvs nano-X the 
: cursor, ie arrow, did not even appear! I checked to make sure I had 
: TUXSCREENMOUSE = Y in config and then after several minutes of fussing 
: and fuming I find that : GdHideCursor(&scrdev);
: had been added to the PD_Open() function in drivers/mou_touchscreen.c. 
: Why, I have no idea.

Hmm.  That was added because the TuxScreen has a touch screen and
by default, showing a cursor wasn't desireable.  So commenting out the
GdHideCursor is the right thing.


: Start the new nano-X and BINGO, I have the cursor. Then my joy is dashed 
: when I touch the screen to move it and check out my fix. It disappears 
: again. I suspect there isd another added GdHideCursor() someplace.

Geez, you're having too much fun trying to get all this too work...  Check
by grepping for GdHideCursor in */*.c, there may be something in 
nanox/srvmain, or in engine/devmouse.c.  I'd check myself but my
main Linux box just seems to have had a HD crash...  (it's also the
one running CVS, but I'll get it all back up shortly)

Regards,

Greg

Greg Haerr wrote:

>: PD_Read() function.
>: 
>:         temp = event.x;
>:         event.x = event.y;
>:         event.y = 1000 - temp;
>: 
>: This now got the cirsor actually following the input device instead of 
>: being 90 degres out of "phase".
>
>Is this code still required in CVS version?  You are running
>in portrait mode, right?  The CVS code automatically swaps
>x/y when in portrait L and R modes.
>  
>
I tried with and without-L or -R and the cusor was always 90 degrees out 
with actual pointer moevment. This is what the pointer movement vs 
cursor movement is,whether landscape or portrait, without my fix.

movement   cursor
    +x            =   +y
     -x             =    -y
     +y            =    -x
      -y            =    +x

With my fix, the cusor moves with pointer in landscape or protrait mode. 
Need to try -D once I get cvs leaving the mouse on the screen. To bad 
the Boss doesn't feel that way.

>: 
>: Cool say I and I move back to the cvs download to achieve the same 
>: thing. But, trouble right off of the bat. When starting cvs nano-X the 
>: cursor, ie arrow, did not even appear! I checked to make sure I had 
>: TUXSCREENMOUSE = Y in config and then after several minutes of fussing 
>: and fuming I find that : GdHideCursor(&scrdev);
>: had been added to the PD_Open() function in drivers/mou_touchscreen.c. 
>: Why, I have no idea.
>
>Hmm.  That was added because the TuxScreen has a touch screen and
>by default, showing a cursor wasn't desireable.  So commenting out the
>GdHideCursor is the right thing.
>
>  
>
I'll look there and see what I find. Actually, looking for all these 
problems has taught me a lot about the code.

>: Start the new nano-X and BINGO, I have the cursor. Then my joy is dashed 
>: when I touch the screen to move it and check out my fix. It disappears 
>: again. I suspect there isd another added GdHideCursor() someplace.
>
>Geez, you're having too much fun trying to get all this too work...  Check
>by grepping for GdHideCursor in */*.c, there may be something in 
>nanox/srvmain, or in engine/devmouse.c.  I'd check myself but my
>main Linux box just seems to have had a HD crash...  (it's also the
>one running CVS, but I'll get it all back up shortly)
>
>Regards,
>
>Greg
>  
>
tj

Another note, 0.90 tarball behaves the same way.

tj

Greg Haerr wrote:

>: PD_Read() function.
>: 
>:         temp = event.x;
>:         event.x = event.y;
>:         event.y = 1000 - temp;
>: 
>: This now got the cirsor actually following the input device instead of 
>: being 90 degres out of "phase".
>
>Is this code still required in CVS version?  You are running
>in portrait mode, right?  The CVS code automatically swaps
>x/y when in portrait L and R modes.
>
>: 
>: Cool say I and I move back to the cvs download to achieve the same 
>: thing. But, trouble right off of the bat. When starting cvs nano-X the 
>: cursor, ie arrow, did not even appear! I checked to make sure I had 
>: TUXSCREENMOUSE = Y in config and then after several minutes of fussing 
>: and fuming I find that : GdHideCursor(&scrdev);
>: had been added to the PD_Open() function in drivers/mou_touchscreen.c. 
>: Why, I have no idea.
>
>Hmm.  That was added because the TuxScreen has a touch screen and
>by default, showing a cursor wasn't desireable.  So commenting out the
>GdHideCursor is the right thing.
>
>
>: Start the new nano-X and BINGO, I have the cursor. Then my joy is dashed 
>: when I touch the screen to move it and check out my fix. It disappears 
>: again. I suspect there isd another added GdHideCursor() someplace.
>
>Geez, you're having too much fun trying to get all this too work...  Check
>by grepping for GdHideCursor in */*.c, there may be something in 
>nanox/srvmain, or in engine/devmouse.c.  I'd check myself but my
>main Linux box just seems to have had a HD crash...  (it's also the
>one running CVS, but I'll get it all back up shortly)
>
>Regards,
>
>Greg
>
>
>  
>