Open forestrf opened 5 years ago
Fix Raycast reporting false positives for rays that cross a bounding box but not the shape that is inside it, thanks to analyzing BEPU Physics raycasting source code.
The code used from BEPU is here: https://github.com/bepu/bepuphysics1/blob/e0438719412e2eab8683d5ca65c1b0bb0e9b177a/BEPUphysics/CollisionTests/CollisionAlgorithms/GJK/GJKToolbox.cs#L347
This should or may fix this issue:
This PR was originally https://github.com/mattleibow/jitterphysics/pull/29 but after removing the repository I encountered this bug https://github.com/isaacs/github/issues/168 so I closed the previous PR to be able to update it. Here are the changes:
After reviewing https://github.com/bulletphysics/bullet3/blob/cdd56e46411527772711da5357c856a90ad9ea67/src/BulletCollision/NarrowPhaseCollision/btSubSimplexConvexCast.cpp I was able to fix the raycast reporting wrong distances randomly by using lambda as fraction directly instead of calculating it.
lambda
fraction
Fix Raycast reporting false positives for rays that cross a bounding box but not the shape that is inside it, thanks to analyzing BEPU Physics raycasting source code.
The code used from BEPU is here: https://github.com/bepu/bepuphysics1/blob/e0438719412e2eab8683d5ca65c1b0bb0e9b177a/BEPUphysics/CollisionTests/CollisionAlgorithms/GJK/GJKToolbox.cs#L347
This should or may fix this issue:
19
This PR was originally https://github.com/mattleibow/jitterphysics/pull/29 but after removing the repository I encountered this bug https://github.com/isaacs/github/issues/168 so I closed the previous PR to be able to update it. Here are the changes:
After reviewing https://github.com/bulletphysics/bullet3/blob/cdd56e46411527772711da5357c856a90ad9ea67/src/BulletCollision/NarrowPhaseCollision/btSubSimplexConvexCast.cpp I was able to fix the raycast reporting wrong distances randomly by using
lambdaasfractiondirectly instead of calculating it.