Hej hej, just want to report that I am using SimplexCollisionShapes and don’t see anything when I am debugging physics and I am curious if SimplexCollisionShapes work or not or it is just debug visualization problem? 
Anyway, I have lots of flat quad objects in my scene. Is it better to use SimplexCollisionShape or MeshCollisionShape for them?
I also try to merge some of these quads visually to save some vertexes. Is it then better to use MeshCollisionShape for merged mesh (which is not necessary quad anymore) or use multiple SimplexCollisionShapes in one CompoundCollisionShape to cover that big mesh.
I am merely interested in performance difference, didn’t have yet the capabilities to test it myself…
It depends on what you are doing. Huge meshes probably perform equally bad as a huge amount of simplex shapes. Generally I’d say mesh prototypes that you combine should work best.
What do you mean with mesh prototypes?
Ok, but if we are talking about only one quad, its definitely better to use SimplexCollisionShape?
Well I doubt you actually see one quad in your game later, so the basic forms you can expect could be mesh prototypes, think tetris or lego.
Thanks for opinions!
Eee, almost forgot, pls take a look at debug visualization of SimplexCollisionShapes, because I still can’t get them to show.
Seems like bullet doesn’t supply one…