[Haskell-cafe] Lazy parser with parsec.

I'm trying to figure out how to write a simple parser in Parsec to tokenize a subset of RTF. The problem is that I haven't been able to come up with a way of writing the parser that doesn't try consuming all of the input just to return the first token. The 'many' primitive's implementation uses an accumulator, and obviously has to parse to the end. Trying to iterate myself causes stack overflows on large inputs. Does anyone know of any existing Parser parsers that don't consume their entire input, or am I probably best off making my own parser. Thanks, David