Apply clang-format-all.sh using the _clang-format file through all the cpp/.h files.
make sure not to apply it to certain serialization structures, since some parser expects the * as part of the name, instead of type.
This commit contains no other changes aside from adding and applying clang-format-all.sh
p.loadURDF("r2d2.urdf", flags=p.URDF_USE_IMPLICIT_CYLINDER)
allow to enable/disable deterministicOverlappingPairs through an API
p.setPhysicsEngineParameter(deterministicOverlappingPairs = False)
Option to de/serialize btPersistentContactManifolds and fix lossy conversion during btMultiBody de/serialization of base world transform
(serialize the quaternion, not the converted 3x3 matrix)
There are still several caches not taken into account, and btMultiBody links/constraints are not deserialized yet etc.
See examples\pybullet\examples\saveRestoreState.py for an example.
For OSX:
cd build
./premake_osx xcode4
for iOS:
cd build
./ios_build.sh
./ios_run.sh
Also integrated the branches/StackAllocation to make it easier to multi-thread collision detection in the near future. It avoids changing the btCollisionObject while performing collision detection.
As this is a large patch, some stuff might be temporarily broken, I'll keep an eye out on issues.
in VoronoiFractureDemo (store intermediate values in memory to avoid problems)
@todo: rewrite the island management to avoid sort for better performance
Thanks to Peter Kyme for the report, fixes Issue 607
It can be disabled by commenting out the #define STATIC_SIMULATION_ISLAND_OPTIMIZATION in Bullet/src/BulletCollision/CollisionDispatch/btUnionFind.h
See also Issue 406
Revert definition for ATTRIBUTE_ALIGNED16, and try to force sizeof(btSolverConstraint) by using unions with btScalar, for non-btScalar data types.
Use btAssert and not assert.
Don't access btAlignedObjectArray elements, for zero sets
Thanks Martijn Reuvers from Two Tribes B.V. (www.twotribes.com) for the patch
To make this work more visible, suppress warnings in external libraries in Extras (COLLADA_DOM, libxml and glui contain many warnings)
Added PreprocessorDefinitions: _CRT_SECURE_NO_DEPRECATE;_CRT_NONSTDC_NO_DEPRECATE to vcproj files
- replaced all usage of heapSort by quickSort
- finished btMultiSapBroadphase. Still some work to be done to increase performance (faster add/remove from child broadphases.
uses currently AABB tree to locate child broadphase (should use grid), and sorted array takes too much time.
This was harder using default arguments. Thanks Vangelis Kokkevis for pointing this out.
- Fixed memoryleak in the ConstraintDemo and Raytracer demo.
- fixed issue with clearing forces/gravity at the end of the stepSimulation, instead of during internalSingleStepSimulation.
Thanks chunky for pointing this out: http://www.bulletphysics.com/Bullet/phpBB3/viewtopic.php?f=9&t=1780
- Disabled additional damping in rigid body by default, but enable it in most demos. Set btRigidBodyConstructionInfo m_additionalDamping to true to enable this.
- Removed obsolete QUICKPROF BEGIN/END_PROFILE, and enabled BT_PROFILE. Profiling is enabled by default (see Bullet/Demos/OpenGL/DemoApplication.cpp how to use this).
User can switch off profiling by enabling define BT_NO_PROFILE in Bullet/src/btQuickprof.h.
Some dynamic memory allocations have been replace by pool allocation or stack allocations.
quantized aabb versus quantized aabb overlap check is made branch-free (helps a lot on consoles PS3/XBox 360)
Collision algorithms are now created through a new btDefaultCollisionConfiguration, to decouple dependency (this is the API change):
Example:
btDefaultCollisionConfiguration* collisionConfiguration = new btDefaultCollisionConfiguration();
m_dispatcher = new btCollisionDispatcher(collisionConfiguration);
CompoundShapes are tricky to manage with respect to persistent contact points and swapped order of btCollisionObjects,
During dispatch, finding an algorith etc. order can be swapped.
fixed several other issues, related to SimpleBroadphase (removing a proxy was not working)
removed deletion of motionstate in CcdPhysics/CcdPhysicsController destructor
disabled m_type optimization in SimdTransform (making the memory size 64 byte, potentially more cache friendly)
fixed a bug in island generation, causing the activation not propagating in one frame, but one 'layer' of rigidbodies at a time
