Lucene 5.4.0 值得关注的改进:
API 改进
* Query.getBoost and Query.setBoost are deprecated in favour of the
new BoostQuery
* The Filter> BooleanQuery
* DefaultSimilarity has been renamed to> prepare for the move to BM25 in Lucene 6
新特性
* New Serbian token filter
* New DecimalDigitFilter, to fold unicode digits to latin digits
* New UnicodeWhitespaceTokenizer, that uses Unicode's whitespace
definition and splits on NBSP
* New GeoPointDistanceRangeQuery to search for geo-points within
a ring
* Query caching is now enabled by default in IndexSearcher, use
IndexSearcher.setQueryCache(null) to disable
优化
* MatchAllDocsQuery got faster
* Doc values now use less memory for multi-valued fields and
less disk in case of sparse fields
* Two-phase iterators got a match cost API so that the costly
bits can be checked last