This parser is implemented using PLY (which is a python implementation of Lex/YACC).
This implementation does not do any error checking. It's possible that it'll succeed parsing erroneous .css files (however, it *should not* fail parsing any valid .css file).
As a package this is worthless. Use Zanderloo, which is fully featured, documented, and totally worth your time.
I've added this basically to just have it stored away, and someone might be interested in the tokenizer/grammar rules.