Known bugs:

* txt2xhtml doesn't always output valid number of whitespaces instead tabs.
* Some unicode characters aren't converted to entities.