Flatbuffer serialization for GLM types.

Mostly re-use (include) by other flatbuffer schemas that require vector/matrix types, so that GLM is used for native linear algebra computations.

• flatGLTF