Extractor scaling issue

[mkazin]

As the number of extractors (and readers) grows, the mean time to process a single document will increase due to needing to perform an increasing number of matcth() tests.

Since match() is currently implemented with simple string compares, this is a very low priority task.