I am attempting to write a new custom viewer for OpenRAVE. As part of this process, I need access to the visual meshes of each kinbody link. As far as I can tell, the qtcoin viewer gets these meshes by loading the files directly, and the visual meshes are simply not stored in OpenRAVE. So, I am limited to rendering the collision meshes.
This is a bit disappointing, as OpenRAVE is somehow loading .IV files to get the collision meshes, but the .IV loader is not accessible (as far as I can tell) from the public API.
What's the best way to access the visual meshes from within OpenRAVE?
I hate to reply to this again, but looking through the qtcoin source code, I think I've realized two things:
1. "GetCollisionMesh" could return the visual mesh, or the collision mesh, depending on which "geometry" is being considered.
2. "Indices" misleadingly does not load an index buffer from the IV files, but rather just stores the numbers 1...n. Vertex data is copied and repeated for each vertex, throwing away the index buffer altogether.