Johannes Kuhlmann

Googling around a little, it seems like the Logitech Dual Action is not directly supported by Xinput, but there are 3rd party drivers/emulators to make it compatible.

Looks like like that wouldn't cause too much pain to remove extensions for those three targets. We could just do it and re-add in case anyone needs it in the...

I've put together my current thoughts on extensions in #125. Feedback would be appreciated.

Did you see the file options in `cgltf_options`? There are `read` and `release` callbacks which should hopefully be sufficient to access data from buffers or a VFS library. They default...

Alright, your points all make sense. I've already got examples on my list. And, yes, it looks like cgltf should resolve reserved URI character encodings. However, about resolving URIs, doesn't...

Have you seen the `cgltf_validate()` function? I'm pretty sure it validates the indices. In my opinion, that's also the better approach because you can verify that the data is valid...

To be honest, I don't really feel the need to add an option to automatically validate. Explicity calling `cgltf_validate()` seems more explicit and more obvious than having an option that...

Would it make sense for this to be optional? I suppose in that case the `char*` would either point to the extra allocation or into the JSON data.

Oh, yes, #114 is kind of a hefty change. I feel it's whole lot of administrative overhead for a use case that's maybe not very common. I'm kind of thinking...

I meant making the copying of the extras data to a new allocation optional. My goal here is, of course, to make everyone happy. So, my preferred way would be...