Event trail - Hacky patch enclosed

Brad Campbell brad at seme.com.au
Thu Aug 23 01:08:36 CDT 2001


Brad Campbell wrote:

> Thanks for the reply Gerard, I can probably handle the traces ;p)
> This is the sort of info I was looking for, I now know where to
> start.
> Am I looking at the best debug output? or is there another
> debug stream that is more suited to what I'm tracing?
> 
> I'll start on it right away.. Thanks..

I have a terrible, kludgy fix that works perfectly.. I'd like to investigate
further why this needs to happen.

--- event.c	Thu Aug 23 14:04:29 2001
+++ event.c.new	Thu Aug 23 14:02:25 2001
@@ -478,6 +478,10 @@
         break;
     }
 
+    X11DRV_SendEvent( MOUSEEVENTF_MOVE | MOUSEEVENTF_ABSOLUTE, pt.x, pt.y,
+                          X11DRV_EVENT_XStateToKeyState( event->state ), 0,
+                          event->time - X11DRV_server_startticks, hWnd);
+
     X11DRV_SendEvent( statusCodes[buttonNum], pt.x, pt.y,
                       keystate, wData, event->time - X11DRV_server_startticks, hWnd);
 }

When a touchscreen buttonpress occurs, it sends NO mousemove events.
If I click the mouse button after doing a touch, it has the same result, but
if I then move the mouse even a tiny bit, it works ok. So I send a mousemove
event at the calculated button coordinates first, and it works fine.
This is a NASTY HACK, and I'll keep looking to see why it is required..
but this does work for now.

-- 
Brad....
                   /"\
Save the Forests   \ /     ASCII RIBBON CAMPAIGN
Burn a Greenie.     X      AGAINST HTML MAIL
                   / \




More information about the wine-users mailing list