Hi Stephen, yes, it's Parsec, but I write this parser only for learning the language, so I don't care if I rewrite existing code. But thanks for the advice regarding performance. Patrick's answer was perfect for me and I will get used to <$> = fmap soon. Tim 2010/11/21 Stephen Tetley :

Where is the Parser monad coming from?

If its Parsec, a "seasoned Haskeller" would avoid read (which is re-parsing already parsed data) and use the integer parser from Parsec's Token parser module.[*]

[*] The one caveat is that the number parsers in this module follow Haskell's lexical rules - you might not always want this. _______________________________________________ Beginners mailing list Beginners@haskell.org http://www.haskell.org/mailman/listinfo/beginners

On 22 November 2010 02:59, Lyndon Maydwell wrote:

I've found that the token library imposes conventions on whitespace that make it difficult to use in many cases.

True - as it can't be made to differentiate between newline and other whitespace it does make writing line oriented parsers difficult. I tend to write my own cut down version of the token modules at that point.