Cube extension improvement

by Stas Kelvich for PostgreSQL Project

In present state cube extension allows to create indices suitable for search queries with several diapasons (e. g. diapasons for price, dates, distance). The main issue of the cube extension is the big index size, therefore building index and reading a large number of records becomes very expensive disk-related operation. More careful leaf node organization in R-Tree can significantly reduce index size. Also with new kNN framework it is possible to create ordering operator for retrieval sorted data directly from the index and ordering operators for kNN with different spatial norms.