support referenced content extension

[andrewheumann]

BACKGROUND:

• In order to support faster function execution, we want to explore skipping GLB merge, and passing GLB urls directly to a client. To do this, we want to support an experimental custom GLTF extension which lets you mark an object as containing content from a referenced URL.

DESCRIPTION:

• Adds support for the experimental referenced content extension. It will not be used unless Elements.Serialization.glTF.GltfExtensions.UseReferencedContentExtension = true; is on.

TESTING:

• I created several functions using this extension and tried them with a custom branch of the hypar front-end that knew how to read the extension.

---

This change is 

[wynged]

mostly looks pretty good, I want to talk about how this could make an easy hop to supporting a glbRepresentation type, see if that desirable, and maybe change some language to support that.

Reviewed all commit messages.
Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @andrewheumann)

---

Elements/src/Serialization/glTF/GltfExtensions.cs line 1223 at r1 (raw file):

                            var nodeId = NodeUtilities.CreateNodeForMesh(meshId, nodes, content.Id, transform);

                            if (contentElement != null && UseReferencedContentExtension)

contentElement != null should already be true at this point.

---

Elements/src/Serialization/glTF/GltfExtensions.cs line 1283 at r1 (raw file):

                    {
                        AddExtension(gltf, nodes[nodeId], "HYPAR_referenced_content", new Dictionary<string, object> {
                            {"contentUrl", contentElement.GltfLocation}

I think we should change the language here away from contentURL towards glbURL.

[andrewheumann]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @wynged)

---

Elements/src/Serialization/glTF/GltfExtensions.cs line 1223 at r1 (raw file):

Previously, wynged (Eric Wassail) wrote…

contentElement != null should already be true at this point.

Done.

---

Elements/src/Serialization/glTF/GltfExtensions.cs line 1283 at r1 (raw file):

Previously, wynged (Eric Wassail) wrote…

I think we should change the language here away from contentURL towards glbURL.

I'm not sure I agree. One reason I don't want to do this is that the format for the referenced content may not always be a GLB. Because we're no longer doing any merging in the function, we can now easily do referenced content of any type threeJS can display: Rhino files, point clouds, objs, etc.

[wynged]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @andrewheumann)

---

Elements/src/Serialization/glTF/GltfExtensions.cs line 1223 at r1 (raw file):

Previously, andrewheumann wrote…

Done.

I don't see it.

---

Elements/src/Serialization/glTF/GltfExtensions.cs line 1283 at r1 (raw file):

Previously, andrewheumann wrote…

I'm not sure I agree. One reason I don't want to do this is that the format for the referenced content may not always be a GLB. Because we're no longer doing any merging in the function, we can now easily do referenced content of any type threeJS can display: Rhino files, point clouds, objs, etc.

yup, I buy that. ContentElements are dead long live Content!

[andrewheumann]

Reviewable status: 1 change requests, 0 of 1 approvals obtained (waiting on @wynged)

---

Elements/src/Serialization/glTF/GltfExtensions.cs line 1223 at r1 (raw file):

Previously, wynged (Eric Wassail) wrote…

I don't see it.

forgot to push oops!

Code quote:

 if (contentElement != null && UseReferencedContentExtension)

[wynged]

Reviewed 1 of 1 files at r2, all commit messages.
Reviewable status: 1 change requests, 0 of 1 approvals obtained

[wynged]

Reviewed 1 of 2 files at r1.
Reviewable status: complete! 1 of 1 approvals obtained