Prev: Remote desktop no longer full screen
Next: Where did windows 7 forum go ????
From: Ryan H. Kawicki Ryan H. on 5 May 2010 14:06
We have a 32 bit process that uses several thread only WH_MOUSE callbacks.
We are in the process of testing the application under 64 bit versions of
Windows. The issue we are seeing has to do with the mouse data value not
being correct after the first call to CallNextHookEx.

Let me start from the beginning.
There are three thread only WH_MOUSE callbacks installed by the application.
When the first callback is notified of mouse activity, I check the lParam
value in memory and notice that either 0x0078 (120) or 0xFF88 (-120) is at
the end of the structure defined for WH_MOUSE callbacks. No modification is
made to the structure of data, so CallNextHookEx is called, which in turn
calls the next installed hook. The lParam is different but the data is the
same, except for the last four bytes of data. It does not contain the wheel
delta value. Same situation happens again when passing this value on through
CallNextHookEx.

I stepped through our application on a 32 bit version of Windows and noticed
that the lParam is changing there as well, so I expect this behavior to
correct, but I expect that the problem lies in what seems to be a copying of
data between calls to CallNextHookEx.

I'm not quite sure where the disconnect is at. So the first time
CallNextHookEx is called in our application, the value is correct, but
subsequent calls are incorrect. This doesn't seem to be the designed
behavior for 32 bit application on 64 bit versions of Windows. Is there
anything in the works to correct this issue? Is there something that we are
doing wrong in our application? Thanks for any help provided.
 | 
Pages: 1
Prev: Remote desktop no longer full screen
Next: Where did windows 7 forum go ????