This updates Cryptomatte to Arnold 5 and does some formatting changes. I've tested transparency and regular usage and it seems to work. I've only tested Maya and haven't tested the user data driven ones yet, but there's no reason to suspect those wouldn't work.
There's another rough spot right now where the pass through shader is required to compute the opacity in the closures, which should probably be handled by cryptomatte.h, but this is perfectly functional for testing. I'd also like to get Solid Angle's eyes on how I'm doing the opacity computation. I don't know how expensive that is or whether that work could be skipped if the opacity flag is off, and whether this would be a good idea from a performance standpoint.
This updates Cryptomatte to Arnold 5 and does some formatting changes. I've tested transparency and regular usage and it seems to work. I've only tested Maya and haven't tested the user data driven ones yet, but there's no reason to suspect those wouldn't work.
There's another rough spot right now where the pass through shader is required to compute the opacity in the closures, which should probably be handled by cryptomatte.h, but this is perfectly functional for testing. I'd also like to get Solid Angle's eyes on how I'm doing the opacity computation. I don't know how expensive that is or whether that work could be skipped if the opacity flag is off, and whether this would be a good idea from a performance standpoint.