Support quad meshes

[GeometryCollective]

Implement basic support for quad meshes by triangulating internally, and copying the solution back to the original quad mesh upon write.

To make this work, tag all "diagonals" inserted by the triangulation, so that cuts never pass through edges that aren't in the original quad mesh.

[JokerMartini]

We started testing this and it appears that the software doesn't successfully import any OBJ mesh other than the one that ships with the application. I attempted to import an OBJ exported from Maya, Max, Zbrush and Mudbox. All of them failed to import.

[GeometryCollective]

Can you try loading one of the meshes in the input/ subdirectory? Someone else we spoke with was also having trouble loading meshes; might be a file I/O problem rather than a mesh parser issue. (The face mesh is hard-coded into the executable, so it doesn't need to do any disk I/O).

Also helpful if you could attach some of the meshes you're trying to load to this thread.

Thanks!

[JokerMartini]

I tried what you suggested and that didn't appear to work either. I've attached a few files for testing. Once of them is a simple Box (1 mesh, 1 poly island) and the Other is a Utah Teapot (1 mesh, 4 poly islands) from 3ds Max.

I also tried making the obj composed of only tri's and that didn't work either. Let me know if i can do anything else to help out.

Box: https://www.dropbox.com/s/3hmx0dxk03b8gzj/box.obj?dl=0 Teapot: https://www.dropbox.com/s/6eso7fhlv36dskt/teapot.obj?dl=0

[GeometryCollective]

Ok, actually several things going on here (listed below). Thanks for being a beta tester—clearly we still have a few kinks to work out in our OBJ support! :-)

• the OBJ files use negative face indices, which is not currently supported by our parser (but should be); I've created a new open issue #8 for this item
• some of the faces in the cube are quads rather than triangles, which is covered by this open issue (#4) but still needs to be implemented
• the teapot consists of multiple components, which is addressed by issue #6 and still needs to be resolved (again this is an easy fix: just map each object separately, but needs to be implemented)
• the teapot also has pieces (like the handle and spout) that have multiple boundary loops; here I've opened an issue #9 to cut them into single disks

In short: there are a lot of little things that need to be fixed, but ultimately shouldn't get in the way of unwrapping meshes like these—see the image below, where I've gone ahead and triangulated/cut the meshes in a modeler, and saved them w/ positive face indices.

[JokerMartini]

Wow this stuff looks great. Seems like you were able to manually get them up and running quickly. I'm not sure how long each of the items will take for you to implement, however as you begin to resolve them, I'd be more than glad to test them out for you. I see a lot of potential in this application and I'm excited to see where it goes.