MarchingCubes::compute_intersection_points is broken

[avaer]

Thanks for this!

I think the code here is busted though. It's using local i j k but I think it should be using the instance vars _i _j _k.

The instance vars are what's actually used in the setting. Also those instance vars won't actually be set at all during construction so this will crash depending on your environment.

I think this would be fixed if we use _i and friends.

[weshoke]

You're right. Presumably this should be changed as well.

[shwilks]

Thanks also for this code! I've corrected the error discussed here, I notice there's no licence in this repo but it would be useful if it's ok with you to incorporate this code into an R package? I would want to make sure that's ok and that I get the code attribution right though, I found it in the first place because this same code is included in scikit for python: https://github.com/scikit-image/scikit-image/blob/master/skimage/measure/mc_meta/MarchingCubes.cpp