- fvi_ray_boundingbox() is faster and should be less prone to compiler/optimization/platform specific issues
- clean up or obj_check_all_collisions() which is a good bit faster now
collideshipweapon.cpp line 262, beam.cpp line 2771, and modelinterp.cpp line 2949.
it needs a better renderer, but is in prety good shape for now,
I also (think) I squashed a bug in the warpmodel code