2 commits

Author	SHA1	Message	Date
Olivier Teulière	01297ebdb4	Sort the word list before processing it. It allows a much better compression, and avoid problems when a word is a prefix of the word just before. ... Also, the word list is not a raw buffer anymore, which makes it a bit easier to understand the algorithm.	2010-05-16 10:06:01 +00:00
Olivier Teulière	9d5c9fe1d0	The dictionary creation is now encapsulated into a dedicated class, to allow reusing it easily	2010-05-15 12:14:53 +00:00