We implement two fulltext parsers:
1. ngram parser for CJK;
2. mecab parser for Japanese.

We use different partition schema for CJK fulltext index tables.

Reviewed-by: Jimmy Yang <jimmy.yang@oracle.com>
Reviewed-by: Marko Makela <marko.makela@oracle.com>
Reviewed-by: Tor Didriksen <tor.didriksen@oracle.com>
RB: 7328