The lexical structure of Haskell is pretty straightforward.
The only problem was that I could not find a good lexer in cpan, so I hacked one together using simple regexes. I am a bit uncertain about what the lexer should return, currently it is the name of the lexeme (or whitechar/comment or ncomment) and the lexeme, this is sufficient for now but probably a bit more would be useful. There is no proper error handling and also there is no tracking of source location (so the parser can not give meaningful error locations). But it works and I can continue with the parser, see how that goes and what kind of requirements that gives for the lexer.