Added KHR_materials_common materials/techniques/shaders to addDefaults

[likangning93]

@JoshuaStorm and I wrote this to have defaultMaterials try to match the existing gltf material values to one of the KHR_materials_common materials, apply the appropriate extension, and then pass the gltf to modelMaterialsCommon at the end of addDefaults.

Also added some tests to addDefaultsSpec.

See #1

[coveralls]

Coverage increased (+1.0%) to 93.274% when pulling 502e3e3 on materialsStageKHR into 3db3306 on master.

[pjcozzi]

@lilleyse can you please do an initial review here?

[coveralls]

Coverage decreased (-5.6%) to 86.717% when pulling 1e68c45 on materialsStageKHR into 3db3306 on master.

[likangning93]

Updated. When testing assimp exported gltf (gltf without techniques) we ran into a problem where cesium's modelMaterialsCommon would produce shaders with empty string uris. We got around it with a customized modelMaterialsCommon in the lib folder, similar to what's in OBJ2GLTF.

Unfortunately assimp exported gltfs still don't seem to work, but the test case we've been using is also weird. It has buffer views for indices, positions, normals, and uvs but they all point at the same buffer and all have offset 0. Very strange. See attached.

assimp.zip

[pjcozzi]

For assimp, have you tried just the box COLLADA model, https://github.com/KhronosGroup/glTF/tree/master/sampleModels/Box/collada

[likangning93]

I gave it a try, but it seems to have the same problem with the bufferViews. It actually runs into the same error in the pipeline and when I try to view it in the online COLLADA to glTF convertor.

[pjcozzi]

What about using a simple OBJ file with assimp, perhaps from https://github.com/SaschaWillems/Vulkan/tree/master/models

Just trying to understand if it is the assimp importer or assimp glTF exporter.

[likangning93]

I tried it with a basic cube and with hammardillo, but it seems to be the same problem, bufferViews that all start at 0.

[likangning93]

I'm pretty sure it's in the assimp gltf exporter. I'm going to try and take a look at their code.

[pjcozzi]

If you figure it out in a reasonable amount of time, open a pull request; otherwise, some an issue to the assimp repo and include sample data and simple instructions to reproduce it.

[lilleyse]

Talked offline, but you use Cesium's defined, defaultValue, and WebGLConstants instead for modelMaterialsCommon.

[likangning93]

Ok. Updated!

[lilleyse]

Nice, looks good!

[coveralls]

Coverage decreased (-5.0%) to 87.336% when pulling ddba592 on materialsStageKHR into 3db3306 on master.

[lilleyse]

@pjcozzi this is ready whenever you want to look

[pjcozzi]

To follow our style guide, just write

return [0, 0, 0, 1];

And drop the else; it is not needed.

[pjcozzi]

Looks like test coverage could be improve a good bit: https://coveralls.io/github/AnalyticalGraphicsInc/gltf-pipeline?branch=materialsStageKHR

[coveralls]

Coverage increased (+1.5%) to 93.815% when pulling 9d7cf7c on materialsStageKHR into 3db3306 on master.

[likangning93]

Updated! Apparently I misunderstood extensions versus extensionsUsed, so fixing that and adding some more to the tests I had helped increase the coverage.

[pjcozzi]

@lilleyse please merge when ready.

[lilleyse]

Looks ready to me.

[pjcozzi]

I deleted the branch. @lilleyse please delete branches after merging just like the Cesium repo.

[lilleyse]

Ah forgot, thanks!