Page MenuHome

Fix T47198: Broken UV/VCol in BI and Cycles with some degenerated faces.
AbandonedPublic

Authored by Bastien Montagne (mont29) on Jan 19 2016, 7:40 PM.

Details

Summary

Problem is actually similar in both engines - in some cases, we changed
'natural' quad splitting order to alternative one, without properkly 'notifying'
UV/VCol/other tessface data about it.

So code would use a 'wrong' triangle of UVs etc.

Solution for BI is rather simple and hopefully does not break anything.

Solution for Cycles is not so elegant imho, but coul not imagine a better way to do it...

Diff Detail

Repository
rB Blender
Branch
t47198

Event Timeline

Sergey Sharybin (sergey) edited edge metadata.

Blender internal side looks fine. Cycles part i was not happy about with all this magic of using negative nverts to indicate some special condition. Added a per-face flag now and added some utility function to get triangles split, so we don't duplicate logic that much.

Those changes are committed at rBa6eae73. Please go ahead with the BI side.

This revision is now accepted and ready to land.Jan 20 2016, 9:17 AM
intern/cycles/blender/blender_mesh.cpp
203

Wrong indices. Proper split is 0, 1, 3, 2, 3, 1 here. Just mentioning.