some doxygen updates, added some comments to classes, fix broken links, rename some prefix some internal GIMPACT collision structures using (for example use GIM_AABB instead of BT_AAABB),

removed obsolete btGjkEpa (it was replaced by btGjkEpa2 ages ago)

src/BulletMultiThreaded/PosixThreadSupport.h

@@ -32,7 +32,7 @@ subject to the following restrictions:
 typedef void (*PosixThreadFunc)(void* userPtr,void* lsMemory);
 typedef void* (*PosixlsMemorySetupFunc)();
 
-// PosixThreadSupport helps to initialize/shutdown libspe2, start/stop SPU tasks and communication
+///The PosixThreadSupport implements the btThreadSupportInterface using pthreads, to help porting SPU Tasks.
 class PosixThreadSupport : public btThreadSupportInterface 
 {
 public:

src/BulletMultiThreaded/SequentialThreadSupport.h

@@ -29,7 +29,8 @@ typedef void* (*SequentiallsMemorySetupFunc)();
 
 
 
-///SequentialThreadSupport is a portable non-parallel implementation of the btThreadSupportInterface
+///The SequentialThreadSupport is a portable non-parallel implementation of the btThreadSupportInterface
+///This is useful for debugging and porting SPU Tasks to other platforms.
 class SequentialThreadSupport : public btThreadSupportInterface 
 {
 public:

src/BulletMultiThreaded/SpuNarrowPhaseCollisionTask/Box.h

@@ -27,7 +27,7 @@ enum FeatureType { F, E, V };
 //----------------------------------------------------------------------------
 // Box
 //----------------------------------------------------------------------------
-
+///The Box is an internal class used by the boxBoxDistance calculation.
 class Box
 {
 public:
@@ -81,6 +81,7 @@ Box::GetAABB(const Matrix3& rotation) const
 // BoxPoint
 //-------------------------------------------------------------------------------------------------
 
+///The BoxPoint class is an internally used class to contain feature information for boxBoxDistance calculation.
 class BoxPoint
 {
 public:

src/BulletMultiThreaded/SpuNarrowPhaseCollisionTask/SpuContactResult.cpp

@@ -139,7 +139,7 @@ bool ManifoldResultAddContactPoint(const btVector3& normalOnBInWorld,
 		newPt.m_combinedRestitution = combinedRestitution;
 
 		/*
-		//potential TODO: SPU callbacks, either immediate (local on the SPU), or deferred
+		///@todo: SPU callbacks, either immediate (local on the SPU), or deferred
 		//User can override friction and/or restitution
 		if (gContactAddedCallback &&
 			//and if either of the two bodies requires custom material

src/BulletMultiThreaded/SpuNarrowPhaseCollisionTask/SpuGatheringCollisionTask.cpp

@@ -266,9 +266,6 @@ public:
 
 		//		spu_printf("processNode with triangleIndex %d\n",triangleIndex);
 
-///TODO: add switch between short int, and int indices, based on indexType
-
-		// ugly solution to support both 16bit and 32bit indices
 		if (m_lsMemPtr->bvhShapeData.gIndexMesh.m_indexType == PHY_SHORT)
 		{
 			unsigned short int* indexBasePtr = (unsigned short int*)(m_lsMemPtr->bvhShapeData.gIndexMesh.m_triangleIndexBase+triangleIndex*m_lsMemPtr->bvhShapeData.gIndexMesh.m_triangleIndexStride);
@@ -403,7 +400,7 @@ void	ProcessConvexConcaveSpuCollision(SpuCollisionPairInput* wuInput, CollisionT
 		int numBatch = subTrees.size();
 		for (int i=0;i<numBatch;)
 		{
-// BEN: TODO - can reorder DMA transfers for less stall
+			//@todo- can reorder DMA transfers for less stall
 			int remaining = subTrees.size() - i;
 			int nextBatch = remaining < MAX_SPU_SUBTREE_HEADERS ? remaining : MAX_SPU_SUBTREE_HEADERS;
 			
@@ -647,7 +644,7 @@ void	handleCollisionPair(SpuCollisionPairInput& collisionPairInput, CollisionTas
 		cellDmaWaitTagStatusAll(DMA_MASK(1) | DMA_MASK(2));
 
 		// Both are compounds, do N^2 CD for now
-		// TODO: add some AABB-based pruning
+		///@todo: add some AABB-based pruning (probably not -> slower)
 	
 		btCompoundShape* spuCompoundShape0 = (btCompoundShape*)collisionShape0Loc;
 		btCompoundShape* spuCompoundShape1 = (btCompoundShape*)collisionShape1Loc;

src/BulletMultiThreaded/SpuSolverTask/SpuParallellSolverTask.cpp

@@ -720,7 +720,8 @@ static void solveConstraint (SpuSolverConstraint& constraint, SpuSolverBody& bod
 
 
 //-- CONSTRAINT SETUP METHODS
-// Compute the jacobian inverse @@TODO: Optimize
+/// Compute the jacobian inverse 
+///@todo: Optimize
 static float computeJacobianInverse (const btRigidBody* rb0, const btRigidBody* rb1,
 							  const btVector3& anchorAinW, const btVector3& anchorBinW, const btVector3& normal)
 {

src/BulletMultiThreaded/SpuSync.h

@@ -31,6 +31,7 @@ Written by: Marten Svanfeldt
 #include <Windows.h>
 #endif
 
+///The btSpinlock is a structure to allow multi-platform synchronization. This allows to port the SPU tasks to other platforms.
 class btSpinlock
 {
 public:
@@ -67,6 +68,7 @@ private:
 //#include <cell/atomic.h>
 #include <cell/sync/mutex.h>
 
+///The btSpinlock is a structure to allow multi-platform synchronization. This allows to port the SPU tasks to other platforms.
 class btSpinlock
 {
 public:

src/BulletMultiThreaded/Win32ThreadSupport.cpp

@@ -26,7 +26,7 @@ subject to the following restrictions:
 
 
 ///The number of threads should be equal to the number of available cores
-///Todo: each worker should be linked to a single core, using SetThreadIdealProcessor.
+///@todo: each worker should be linked to a single core, using SetThreadIdealProcessor.
 
 ///Win32ThreadSupport helps to initialize/shutdown libspe2, start/stop SPU tasks and communication
 ///Setup and initialize SPU/CELL/Libspe2