Add support for EXT_accessor_additional_types

[lilleyse]

EXT_accessor_additional_types defines additional accessor component types INT, INT64, UNSIGNED_INT64, and DOUBLE.

We're using this branch for some internal pipelines where it's helpful to have higher precision types when manipulating glTF data. The changes really only matter when constructing AccessorViews directly and shouldn't affect the runtimes.

The extension also defines HALF_FLOAT but since std::float16_t is only available in C++23 it's not included here.

[javagl]

My gut feeling is that the spin-off extension at KhronosGroup/glTF#2397 is more likely to become "official" - exactly because it is more fine-grained and avoids the potential hassle of "half" often not being present.

But on the level of the implementation, this should be a true subset of what is now supported here (so... this comment may just be nitpicking/noise 😶 )

[kring]

Thanks @lilleyse!