I think in the simple subdivision routine, the vertex attribute computation somehow got a bit jumbled.
There is some linear triangle-edge interpolation going on - but the new attributes do not match the values that I would have expected.
Thus, I have simplified the per-edge attribute interpolation to be analog to the actual subdiv-vertex position computation and added a testcase to verify.
Hey Mike,
I think in the simple subdivision routine, the vertex attribute computation somehow got a bit jumbled. There is some linear triangle-edge interpolation going on - but the new attributes do not match the values that I would have expected.
Thus, I have simplified the per-edge attribute interpolation to be analog to the actual subdiv-vertex position computation and added a testcase to verify.
Cheers Dennis