Model created without material

[pjcozzi]

@fabrobinet I will send you the COLLADA model.

A simple skinned COLLADA model exported from blender is converted with collada2gltf without warning, but it does not generate any shaders nor does it include a material:

    "materials": {},
    "meshes": {
        "Cube-mesh": {
            "name": "Cube",
            "primitives": [
                {
                    "attributes": {
                        "JOINT": "attribute_40",
                        "NORMAL": "attribute_20",
                        "POSITION": "attribute_18",
                        "WEIGHT": "attribute_37"
                    },
                    "indices": "indices_16",
                    "material": "",
                    "primitive": 4
                }
            ]
        }
    },

Even if this is an issue with the Blender exporter, we need to handle this reasonably.

[antont]

We got a scene from Blender to glTF viewer successfully yesterday (after some hardship).

You can see it live in http://playsign.tklapp.com:8000/glTF-webgl-viewer/ and the source is in http://www.ee.oulu.fi/~antont/download/oulu3dlive-lofi.zip (contains both the blends and the aggregating Masterscene.blend and the original Ogre3D exports).

We needed to set the texture export options so that both 'use uv textures' and 'use material textures' are on, but the 'use only active uv layer' is not checked.

I do agree that it would be nice to not crash when the TEXCOORD param is missing :) (we were first using a python snippet to find meshes without it and removed them from the scene json manually)

[pjcozzi]

Ah, thanks. We'll have a look.

[fabrobinet]

That's a dup of #154 but this new issue got more info, so we'll keep it.

[fabrobinet]

So, we now create a default material (white) for scene coming without material... Actually, 2 kinds of materials/technique/shaders depending if normals are present. Ideally, we'd also like to generate normals when default lighting is turned on for these models. This is tracked more generally here: #202

[Digicrat]

The latest dev-4 branch you committed to resolves my issue CesiumGS/cesium#1385 (and brings me to a new one which is irrelevant to this ticket).