Closed GoogleCodeExporter closed 9 years ago
If you try to attempt to compute the inertia tensor for a
btBvhTriangleMeshShape , you will get an assert. It should be made even more
obvious that btBvhTriangleMeshShape should only be used for static, non-moving
collision geometry.
I added some more comments in the header file, and will add more obvious
information in the Bullet user manual soon (the pdf version, included in the
SDK).
http://code.google.com/p/bullet/source/diff?spec=svn2630&r=2630&format=side&path
=/trunk/src/BulletCollision/CollisionShapes/btBvhTriangleMeshShape.h
For moving concave triangle meshes, it is recommended to perform convex
decomposition, using HACD, see Bullet/Demos/ConvexDecompositionDemo.
Alternatively you can use GImpact for moving concave meshes, but that is not
reliable in all cases, so convex decomposition + btCompoundShape is the only
option.
I'm looking into fixing the gimpact shape, and possibly merge its functionality
into btBvhTriangleMeshShape in the future.
Original comment by erwin.coumans
on 14 Dec 2012 at 8:36
Thanks for the quick response. I was just using bullet for collision checking
so naturally I didn't hit the assert.
Original comment by john.d.s...@gmail.com
on 15 Dec 2012 at 12:48
If you really want moving concave triangle meshes, you could try GImpact. See
Bullet/Demos/MovingConcaveDemo or Bullet/Demos/GImpactDemo
Also, I just created a new utility that can create a btCompoundShape using a
GImpact mesh.
See https://code.google.com/p/bullet/source/detail?r=2631 and
btCompoundFromGImpact.h header
Original comment by erwin.coumans
on 15 Dec 2012 at 1:06
Original issue reported on code.google.com by
john.d.s...@gmail.com
on 14 Dec 2012 at 3:22Attachments: