I have a question about the relevance of collision shape margins in collision detection.
I only use the collision world and collision objects with no dynamics - no rigid/soft bodies, I update object positioning from outside information and then check if there are collisions in the world. Can I set each collision shape margin to 0 without this affecting collision detection reliability or performance?
After looking around on the Bullet forums and documentation, my impression is that the shape margins are used to improve the response to collisions (the actual physics part) rather than the detection itself. Is this impression correct?
Thank you for any information!