Using Parsec for lexing&parsing

Question

I'm creating list of Tokens from input [Char] stream using Parsec v3. The definition of Token looks like this:

data Token = CharKeyword | OpeningBracket | Identifier String | Natural Int

As result of calling parse lexComb "" inputStream i have [Token]. Everything OK so far.

Now i want to parse this list of tokens using another parse invocation, but i've found myself writing such boilerplate code for every Token constructor:

psOpBracket = tokenPrim s np p
    where
        s _ = "'{'"
        p (OpeningBracket) = Just Nothing
        p _ = Nothing
        np pos _ _  = pos

another example:

psPacketName = tokenPrim s np p
    where
        s x = x
        p (Identifier x) = Just x
        p _ = Nothing
        np pos _ _  = pos

I wonder if it's right to use tokenPrim at all. And if so, is there any way to automatically generate parsers for every token? Maybe Template Haskell?

Answer 1

I've solved it by using Text.ParserCombinators.Parsec.GeneralizedToken from MissingH package:

psOpBracket = tokeng (\x -> case x of TokOpBracket -> Just Nothing ; _ -> Nothing)

I had to use tokenPrim for cases when Token constructor takes arguments, though.