This is referring to Mitch's new SkeletonAnimator runtime, which is super cool but has some behavior differences from the older runtime.
I'm trying to implement a depth-based masking feature that is compatible with spine animations. I basically just ZWrite to the buffer an area with my own shader, then ZTest GEqual with a copy of the SpineSkeleton shader at a higher queue order (I've pointed all the spine materials to my own customized shader).
This approach seems to work with normal SkeletonAnimation, but doesn't with SkeletonAnimator. Looking through the code, nothing leaps out as problematic. So, is there some reason you know of that SkeletonAnimator isn't compatible with this approach? I am definitely no expert at Cg, I could easily be missing something.
I can always do a multiple rendertarget approach, compositing the final image as an image mask, but that wouldn't be ideal (mostly for performance reasons).
Thanks!
Jeremy
