Seeks to "largest" indexed term that's <= term; returns file pointer index (into the main terms index file) for that term
Only implemented if TermsIndexReaderBase.supportsOrd returns true.