The output from CreoleParser-3k has a lot of superfluous characters added to it's output that do not appear when using the original CreoleParser. A " b' " mark is added to the top of the document, and all of the newlines are appearing within the generated page text.
This occurs both with a custom dialect parser and with the basic text2html parser.
I did some testing, and apparently it is returning a byte object instead of a string object. In order to work around the issue, I have to pass it through str(output, "utf-8").