/home/travis/build/bulletphysics/bullet3/src/LinearMath/btQuickprof.cpp:251:19: error: ‘CLOCK_REALTIME’ was not declared in this scope
/home/travis/build/bulletphysics/bullet3/src/LinearMath/btQuickprof.cpp:251:37: error: ‘clock_gettime’ was not declared in this scope
Make btQuickprof thread safe
Add option in btQuickprof to override custom timing profile (btSetCustomEnterProfileZoneFunc, btSetCustomLeaveProfileZoneFunc)
remove b3Printf in a user/physics thread (those added added, while drawing the GUI running in the main thread)
(for example solverInfo().m_leastSquaresResidualThreshold = 1e-7 and use large m_numSolverIterations
disable sphere-sphere contact cache, it is buggy (some contact point stay in the cache, when sphere penetrates more than total margins)
tweak some gpu demo settings
allow 'useMaximalCoordinates' and 'useFixedBase' in pybullet.loadURDF.
enable split impulse for btRigidBody, even in btMultiBodyDynamicsWorld.
allow initialization of velocity and apply force for btRigidBody in pybullet/shared memory API.
process contact parameters in URDF also for btRigidBody (friction, restitution etc)
add pybullet.setPhysicsEngineParameter with numSolverIterations, useSplitImpulse etc.
add option to perform filtering of 'getClosestPoints' using linkA/linkB.
don't use 'realtimesimulation' as default
add/remove debug items within same thread
pybullet, report contact points and normal as [x,y,z] triplet/vector, not 3 scalars
separate 'getClosestPointsAlgorithm': box-box doesn't report closest points with positive distance, and the query shouldn't impact regular 'closesst points'
- fixing various race conditions throughout (usage of static vars, etc)
- addition of a few lightweight mutexes (which are compiled out by default)
- slight code rearrangement in discreteDynamicsWorld to facilitate multithreading
- PoolAllocator::allocate() can now be called when pool is full without
crashing (null pointer returned)
- PoolAllocator allocate and freeMemory, are OPTIONALLY threadsafe
(default is un-threadsafe)
- CollisionDispatcher no longer checks if the pool allocator is full
before calling allocate(), instead it just calls allocate() and
checks if the return is null -- this avoids a race condition
- SequentialImpulseConstraintSolver OPTIONALLY uses different logic in
getOrInitSolverBody() to avoid a race condition with kinematic bodies
- addition of 2 classes which together allow simulation islands to be run
in parallel:
- btSimulationIslandManagerMt
- btDiscreteDynamicsWorldMt
- MultiThreadedDemo example in the example browser demonstrating use of
OpenMP, Microsoft PPL, and Intel TBB
- use multithreading for other demos
- benchmark demo: add parallel raycasting
use btAssert to detect 0 determinant in btMatrix3x3 inverse
Remove obsolete comment in PosixSharedMemory, the Windows shared memory implementation was done over a year ago
