The parser unexpected msg always fails with an unexpected error message msg without consuming any input.

The parsers fail, (<?>) and unexpected are the three parsers used to generate error messages. Of these, only (<?>) is commonly used. For an example of the use of unexpected, see the definition of Text.Parsec.Combinator.notFollowedBy.

The parser p ? msg behaves as parser p, but whenever the parser p fails without consuming any input, it replaces expect error messages with the expect error message msg.

This is normally used at the end of a set alternatives where we want to return an error message in terms of a higher level construct rather than returning all possible characters. For example, if the expr parser from the try example would fail, the error message is: '...: expecting expression'. Without the (<?>) combinator, the message would be like '...: expecting "let" or letter', which is less friendly.

This combinator implements choice. The parser p <|> q first applies p. If it succeeds, the value of p is returned. If p fails without consuming any input, parser q is tried. This combinator is defined equal to the mplus member of the MonadPlus class and the (Control.Applicative.<|>) member of Control.Applicative.Alternative.

The parser is called predictive since q is only tried when parser p didn't consume any input (i.e.. the look ahead is 1). This non-backtracking behaviour allows for both an efficient implementation of the parser combinators and the generation of good error messages.

The parser try p behaves like parser p, except that it pretends that it hasn't consumed any input when an error occurs.

This combinator is used whenever arbitrary look ahead is needed. Since it pretends that it hasn't consumed any input when p fails, the (<|>) combinator will try its second alternative even when the first parser failed while consuming input.

The try combinator can for example be used to distinguish identifiers and reserved words. Both reserved words and identifiers are a sequence of letters. Whenever we expect a certain reserved word where we can also expect an identifier we have to use the try combinator. Suppose we write:

  expr        = letExpr <|> identifier <?> "expression"

  letExpr     = do{ string "let"; ... }
  identifier  = many1 letter

If the user writes "lexical", the parser fails with: unexpected 'x', expecting 't' in "let". Indeed, since the (<|>) combinator only tries alternatives when the first alternative hasn't consumed input, the identifier parser is never tried (because the prefix "le" of the string "let" parser is already consumed). The right behaviour can be obtained by adding the try combinator:

  expr        = letExpr <|> identifier <?> "expression"

  letExpr     = do{ try (string "let"); ... }
  identifier  = many1 letter

many p applies the parser p zero or more times. Returns a list of the returned values of p.

  identifier  = do{ c  <- letter
                  ; cs <- many (alphaNum <|> char '_')
                  ; return (c:cs)
                  }

skipMany p applies the parser p zero or more times, skipping its result.

  spaces  = skipMany space

The most general way to run a parser over the Identity monad. runParser p state filePath input runs parser p on the input list of tokens input, obtained from source filePath with the initial user state st. The filePath is only used in error messages and may be the empty string. Returns either a ParseError (Left) or a value of type a (Right).

  parseFromFile p fname
    = do{ input <- readFile fname
        ; return (runParser p () fname input)
        }

parse p filePath input runs a parser p over Identity without user state. The filePath is only used in error messages and may be the empty string. Returns either a ParseError (Left) or a value of type a (Right).

  main    = case (parse numbers "" "11, 2, 43") of
             Left err  -> print err
             Right xs  -> print (sum xs)

  numbers = commaSep integer

updateState f applies function f to the user state. Suppose that we want to count identifiers in a source, we could use the user state as:

  expr  = do{ x <- identifier
            ; updateState (+1)
            ; return (Id x)
            }