Description
I noticed a defect in GltfLoader while developing the MonkeyWrench library.
The test data is TextureTransformMultiTest from Khronos's glTF sample models.
Here's how it looks in CompareLoaders when "TextureTransformMultiTest/glTF/TextureTransformMultiTest" is loaded using jme3-plugins:
The 6 empty boxes at the top of the "UV 1" column indicate that jme3-plugins doesn't correctly process the "texCoord" : 1 lines in "TextureTransformMultiTest.gltf" that select the secondary U-V channel. jme3-plugins is already modifying texture coordinates for the "KHR_texture_transform" extension, so selecting a U-V channel ought to be possible.
(Currently, JME's PBR material does not implement clearcoat,which is why the 9 bottom boxes are completely dark. That's a separate issue!)