http://bitbucket.org/lindenlab/trunk/src/e2726e788831/indra/viewer/llface.cpp

I don’t know C++ but it looks fairly comprehensible ?

As I mentioned, if I had to do it again, I would eliminate the Face concept and just begin with simple projections of the same texture everywhere, applying special cases as decals.
The popping you’re seeing could be from some mismatch between faces and actual sides — maybe some optimization kicking in.

Sorry I couldn’t be of more help.

I am stuck on one single type – that is planar texture applied to the sides of a tapered or sheared box prim – it has to be planar to avoid texture tearing – but there is a shift in the planar numbers when the prim is tapered or sheared – the sloping side texture is offset and stretched. This needs to be corrected and it would be great to get the script to do it – but I can’t figure out how to measure the effect – it looks as thought the planar map is projected from a central point somwhere and distorts when the prim changes shape or proportions. Would love to understand this.

{PS – how I wish I had known about this site a year ago – great stuff

I did have a question as to why the ability to use more of the points isn’t available to the user? I enjoy making complex geometry, and would prefer making it in SL rather then sculpting it.

This would include a more practice (in my opinion) though more complex extrusion system. That would be more like you would find in 3D design software.

For 3, you get different texturing results depending on how you select paths and profiles for making spheres. There’s more than one choice. Convolving a semicircular profile around a circular path gives spherical coordinates. But you could also imagine a sphere built as a circle convoled over a straight line path that has its scale set such that the result is a sphere instead of a cylinder. That could be made to have a much better texture projection, especially for the poles.

Perhaps you could offer some insight on a few things that I have been curious about:

1) It feels to me like Path Cut and Profile Cut are reversed for Cylinder/Box/Prism compared to Torus/Tube/Ring. That is, for Box et al., “Path Cut” seems to clearly edit the profile, while “Profile Cut” seems to edit the path (using a modified client or the dimpled-sphere-to-box trick to edit that attribute). Is this part of the interface, or part of the volume algorithm? And, was this a design decision, an oversight, or am I just thinking about the geometry in the wrong way?

2) Am I correct in thinking that it should be possible to have torus-like prims, but with a square or triangular path instead of a circle? I.e. something visually similar to 4 wooden rods joined into a square frame (or 3 into a triangular frame). Is there support in the volume algorithm for such square and triangle paths?

3) Could you talk about the “Planar Mapping” texture coordinate algorithm? I’m particularly interested in knowing the origins of the strange effect it has when applied to spheres and other rounded prims.