In some degenerate cases the contact normal is pointing the wrong direction
so fix it now (until we can deal with all degenerate cases in GJK and EPA)
contact normals need to point from B to A in all cases, so we can simply check if the contact normal really points from B to A
We like to use a dot product of the normal against the difference of the centroids,
once the centroid is available in the API
until then we use the center of the aabb to approximate the centroid
Added initial support for polyhedral contact clipping.
This clipping takes a separating normal, that can be computed using either SAT or GJK/EPA.
To enable clipping, use btPolyhedralConvexShape::initializePolyhedralFeatures(); (needs to be enabled for both convex shapes)
No concave trimesh support for SAT/clipping yet. To enable SAT, see the toggle in btConvexConvexAlgorithm.
Fixes in contact normal in btGjkPairDetector. Hopefully this doesn't cause any regression (we need unit tests!)
Make btTypedConstraint and btPersistentManifold both derive from btTypedObject to make SPU-side generic constraint solver easier.
Note: all build systems need to be updated: remove SpuVoronoiSimplexSolver.cpp, SpuGjkPairDetector.cpp, SpuEpaPenetrationDepthSolver.cpp, SpuGjkEpa2.cpp
+ Allow user to enable useConvexConservativeDistanceUtil . Use dynamicsWorld->getDispatchInfo().m_useConvexConservativeDistanceUtil = true;
(see Demos/Benchmarks/Benchmark4 (convex objects falling down)
+ Fix for plane drawing (just wire-frame)
+ Gimpact: use collision margin of 0.07 for demo (because BULLET_TRIANGLE_COLLISION is used)
+ replace dot,cross,distance,angle,triple in btVector3 by btDot, btCross,btDistance,btAngle,btDistance to avoid naming conflicts
+ Some fixes in GJK penetration depth normal direction (broken in a previous commit)
+ fix in calculateDiffAxisAngleQuaternion to make ConvexConservativeDistanceUtil work properly
+ allow debug drawing to debug btContinuousConvexCollision
+ add comment/warning that btTriangleMesh::findOrAddVertex is an internal method, users should use addTriangle instead
Thanks Ole for the report and suggested fix:
http://code.google.com/p/bullet/issues/detail?id=220&can=1&start=200
Applied Character/btKinematicCharacterController.cpp patch:
Thanks Thomas for the patch!
http://code.google.com/p/bullet/issues/detail?id=196#makechanges
- added a normalizedDirection member variable. This is
slightly more efficient than recalculating the normalized
direction every simulation step.
- I overloaded the m_walkDirection member variable to hold
either the walkDirection or the Velocity, and used a boolean
flag to determine which behavior was required.
- The normalization behavior seemed custom, and I needed it
twice, so I moved it to a static helper method.
- I added the setVelocityForTimeInterval() method to the
base interface class. This to me is the scariest change
since people inheriting from it will need to make changes
(if only to add an empty function).
Thanks to Ole for reporting
Fix a bug , causing some deep penetrations to be missed. Not likely to have impacted many users, as such deep penetrations should have prevented in the first place.
Thanks a lot to Andy O'Neil for reporting and reproduction case.
BulletMultiThreaded/NarrowPhaseCollision makes use of this boxBoxDistance.
Cache some values in src/BulletMultiThreaded/SpuContactManifoldCollisionAlgorithm.cpp, to avoid DMA transfers
2) Added btConvexSeparatingDistanceUtil: this allows caching of separating distance/vector as early-out to avoid convex-convex collision detection.
btConvexSeparatingDistanceUtil is used in src/BulletCollision/CollisionDispatch/btConvexConvexAlgorithm.cpp and can be controlled by btDispatcherInfo.m_useConvexConservativeDistanceUtil/m_convexConservativeDistanceThreshold
3) Use BulletMultiThreaded/vectormath/scalar/cpp/vectormath/scalar/cpp/vectormath_aos.h as fallback for non-PlayStation 3 Cell SPU/PPU platforms (used by boxBoxDistance).
Note there are other implementations in Extras/vectormath folder, that are potentially faster for IBM Cell SDK 3.0 SPU (libspe2)
+ normalize plane normal input for btStaticPlaneShape
+ fixed issue related to swapped collision detectors (SphereTriangleDetector in particular)
Thanks a lot to Andrey Tuganov for reporting the issue and his reproduction case ( http://www.bulletphysics.com/Bullet/phpBB3/viewtopic.php?f=9&t=2143 )
Added btStackAlloc to Bullet (right now only used by btGjkEpa)
removed default constructors of btCollisionWorld/btDiscreteDynamicsWorld, to reduce link-time dependencies
improved performance of constraint solver by precalculating the cross product/impulse arm
added collision comparison code: ODE box-box, also sphere-triangle
added safety check into GJK, and an assert for AABB's that are very large
write partid/triangle index outside of GJK
