So we can implement lexer and parser automaton compression according to [here](https://github.com/cenotelie/hime/issues/85)
