increase geometric flexibility of regions

[Peter Powers]

UCERF3 led to the development of algorithms for spatial analysis that were difficult to implement using regions because of the singular polygon requirement. Moreover, it appears that the singularity requirement can be broken using the static CAG (union, intersect) methods.

Internally, Regions use java.awt.geom.Areas and they are actually pretty messy leaving lots of 0-area polygons behind and boxers with lots of duplicate points after performing CAG ops. Cleanup methods were developed in UCERF3 that would allow Regions to be more complex and I don't see any good reason not to allow them to be.