When converting a cvf::Vec3d to an integer vector, the platform int type was used as an intermediate variable. This caused overflow for large double values.