Matt has had two of his pull requests in the queue for a while. I think both are very worthwhile changes that will unblock other work, so I'd like to request some more pairs of eyes on them:

The first implements boolean objects, an oft-requested feature from yt 2.x. The second makes it so our pixelization routines produce images via compositing onto an existing image rather than returning an image which does not exist yet. It also fixes some long-standing issues with the orientation of images returned by the pixelizer.

The latter in particular will unblock work on supporting datasets that contain multiple disjoint but continuous meshes - something that's common in numerical relativity data.

-Nathan