Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[lathe]

This fixed an issue I was having at any framerate below 500.  Nice fix, spent a lil too much time hacking my own solution together instead of searching harder

[Svenstaro]

Has this patch ever been merged?
https://sourceforge.net/tracker/index.php?func=detail&aid=1981254&group_id=149835&atid=775955 still shows open. It's been almost two years. What's up? Is OIS not actively developed or maintained anymore?

[Entelin]

The change has not been merged, and the bug still exists in the svn version. The distance the mouse travels on the screen given a physical motion should not change, regardless of fps. The current code in LinuxMouse::_processXEvents() iterates over the entire mouse event stack, however the stack only holds the most current relative motion event and not the sum of everything thats happened, so the lower the fps, the less actual motion is captured.

The fix is to move everything within   if( event.type == MotionNotify ) {....}  out of the while loop and use  if ( XQueryPointer(display, window, &u1, &current_win, &u2, &u2, &x, &y, &mmask) )  to query the mouses state every frame instead, this is what SDL does in their code for example. The patch submitted to the bugtracker earlier in this thread shows what I did exactly.

Its a fairly serious bug... hopefully it will be merged at some point.

[Entelin]

I think my previous patch resulted in mMoved being true all the time,  mMoved = true  needs to be relocated here:

Code: [Select]

//Compute this frames Relative X & Y motion
mState.X.rel = x - oldXMouseX;
mState.Y.rel = y - oldXMouseY;
if (x != oldXMouseX || y != oldXMouseY) mMoved = true;

[pjcast]

If there is a correction to the patch, please update the patch tracker. I hopefully will get to clearing out all patches soon.