First, I'm on an updated yt 3.0 updated about 5 minutes ago. (hg id: ea182fc68fae (yt-3.0) tip)

This might be related to Matt's update on octree on June 8th (in response to the CIC stuff) that I don't quite understand, but I've run into the following issue highlighted in the example code:

http://paste.yt-project.org/show/4781/

Here, I create a fake data set (motivated by Nathan's shared notebook earlier this week), construct the octree, and then smooth the density onto a grid with:

>density = ad["deposit","all_density"]

What I find is that the number of elements in the density array is not the same as the number of 0's (i.e Falses) in the generated octree (revealed by 

>saved["octree"] 

in the above script).   Nominally, I thought the smoothing happened onto the same octree mesh that I generate with ds.index.oct_handler.save_octree().  Is this not the case?  Or, is there a way to make it the case in the order that I call things?

Thanks,
desika