Here's abit more of a demonstration as to why STRIP isn't quite up to snuff. Before proceeding it might be a good idea to brush up on
Alias-Wavefront format. We've got some minor custom extensions going on in our files, but for the most part it's the same.
Here's our base model. It has no UVs, no Vertex Normals, and no Vertex Colors. It looks like this:
Un-stripified OBJ file
stripified version
This, predictably, works just fine with our current importer.
However, when we move onto something a little more complicated, like this model with UV coordinates (aka texture verts) and Vertex Normals:
Un-stripified OBJ file
stripified version
STRIPE makes a fatal assumption. The number of UVs in a model is not guaranteed to be the same as the number of verts in a model. In fact, for that to happen is very rare. Each vertex can have an infinite number of UV coordinates, since the number of uv coordinates should be faces * 3 (before stripifying). I think I know why they made this assumption though, and that is so that they could keep their strip vertex references as a single value, unlike how wavefront's face vertex references can have multiple values (one for each kind of vertex - you'll see them separated by slashes like so: v/vt/vn). Because of this improper handling of UV coordinates, textured models don't work properly. STRIPE even went so far as to include the vertex normals (although, it never references them properly to do the issue mentioned previously) The solution to this would be to add the ability to add the slashes for extra references into the strip and strip continuation entries (t and q respectively).
There's one more case: A model with faces, vertex normals, UV coordinates, and vertex colors. This requires uses of our format extension (we added vc indeces right after the vn indeces in the face definition).
However, when we move onto something a little more complicated, like this model with UV coordinates (aka texture verts):
Un-stripified OBJ file
stripified version
STRIPE tries with this model and doesn't crash. I'm not sure if it imports into sadxlvl2 without crashing the program, but either way, it ignored our vcolors (obviously), and didn't reference the included data properly.