Real-Time Physics Simulation Forum

possible bug

2006-06-21T05:31:08+00:00

Awsome, worked a treat, thanks!

Statistics: Posted by mirv — Wed Jun 21, 2006 5:31 am

possible bug

2006-06-21T04:50:19+00:00

The demo doesn't prevent from picking the ground object.

When picking the ground object, it tries to create a point to point constraint between the ground and the world. Obviously, you won't be able to move the ground, the jacobian elements can't be calculated for fixed objects.

You can fix it by changing line 820 in CcdPhysicsDemo
if (body && !body->IsStatic())

Thanks,
Erwin

Statistics: Posted by Erwin Coumans — Wed Jun 21, 2006 4:50 am

possible bug

2006-06-21T03:10:38+00:00

Hi all,
I've been poking around at the latest release, and noticed that in SimdVector3.h , line 186
return SimdVector3(v1.x() * v2.x(), v1.y() * v2.y(), v1.z() *+ v2.z());
shouldn't that be:
return SimdVector3(v1.x() * v2.x(), v1.y() * v2.y(), v1.z() * v2.z());
?
I've been getting some runtime errors with the demos, and I think that's the main culprit. Testing now anyway.

Edit - still some runtime errors, only seem to occur when I "mash" a box into the ground (with ccdphysicsdemo for example). Call stack gives the same location, looking into it still.

Edit2 - ok, sorry, scratch that, it doesn't seem to affect release builds, only debug. Only reason I can think of is that inlining is typically different in debug builds (I'm using visual c++ express).

Statistics: Posted by mirv — Wed Jun 21, 2006 3:10 am