Closed lilleyse closed 6 years ago
@hpinkos could you check this out? The new code starts here: https://github.com/AnalyticalGraphicsInc/obj2gltf/pull/157/files#diff-608b0b654865416bb99fb5d336cf3307R297
I don't have write access to this repo so I can't merge it, but this looks good to me @lilleyse
Thanks @hpinkos. I'm going to go ahead and merge.
Fixes https://github.com/AnalyticalGraphicsInc/obj2gltf/issues/147 by updating to the latest triangulation code used in Cesium. This fixes a model that wasn't triangulating correctly in
master
. This fix required updating Cesium so I updated other dependencies with it.csxirc5h.i2b.obj
Before and after:
There are still some back facing polygons but I confirmed they are present in the original model too. We don't yet do that sort of correction.
(Backfacing polygons on the steps)
I checked that the new method still triangulates these models correctly:
https://github.com/AnalyticalGraphicsInc/obj2gltf/issues/33 https://github.com/AnalyticalGraphicsInc/obj2gltf/issues/50 https://github.com/AnalyticalGraphicsInc/obj2gltf/issues/79
One thing I'm not completely sure about is whether
PolygonPipeline.computeWindingOrder2D
is enough to check the winding order. Would it need to use the normal like the 3-sided case?