Thanks for the excellent reply!
If I try to make a grid that’s 2xN grid, it changes my 2 back to a 3. Dunno why.
But any how there might be a bug with how DupliFaces handles non-manifold meshes if complex meshes are involved … I first tried with a simple cylinder and extruded some faces after the fact … no problems … then I tried with the suzanne monkey head in the place of the cylinder and the same weird axis rotation (starting from the point where I had extruded new faces) problem you have … I even unparented and saved and reloaded and tried again but the same results …
So maybe you should delete and save and reload then recreate the mesh strip from scratch if you didn’t in the first place … apparently there is some mesh data that affects the orientation of duplifaced objects that end users are not aware of (definitely not an issue with normals) …
Yeah, I was trying to figure out what on earth might be stored in the mesh that would be doing this. Unfortunately, because Blender uses a really obfuscated binary file format, I can’t realistically go snooping in the .blend files for clues.
But could I ask why you are using DupliFaces in your project ? You could just as well use the Array Modifier instead and use the Curve Modifier in conjunction with Hooks to animate your mesh(es) just as well (if not better) … maybe you missed ZanQdo’s (I think that’s his Nom du Blend) demos of the above modifiers ? Using the combination of the above modifiers is ideal for vertebra/spinal columns … Personally I found the entire “dupli” family of modeling a bit buggy when I tried it …
So, I did think about using the array modifier. I actually experimented a little with the array modifier, with a repeating cube, when I was getting frustrated with the duplifaces problem. I didn’t try it with a curve; I used an armature, instead. It distorted the heck out of the array items – since it was deforming them as though they were one mesh, rather than treating them as articulated pieces. I don’t want my vertebrae stretching and twisting all over the place.
Just now, I’ve tried it with a curve, and I’m still seeing distortion in my cubes as the curve angle gets sharper. Given that the vertebrae are fairly flat, the ribs (which will ultimately be animated in similar fashion) are already tapered, and that the spine is ultimately limited in its flexibility, it may be an acceptible amount of distortion.
I’ll do a test with it, to see if I can control it adequately. My tests with the cubes have me a little worried about my meshes sliding around all over the place.