Haskell Language Learning Note (parsec) (3)

Applicative parsing

Use the applicative type of Parser.
Operators include use (<$>), (<*>), (<$), (<*), (*>), (<|>), many, and so on.

import Control.Monad-- showimport Text.Parsecimport Control.Applicative hiding ((<|>))number = many1 digitplus = char '+' *> numberminus = (:) <$> char '-' <*> numberinteger = plus <|> minus <|> numberfloat = fmap rd $ (++) <$> integer <*> decimal    where rd      = read :: String -> Float          decimal = option "" $ (:) <$> char '.' <*> number-- /showmain = forever $ do putStrLn "Enter a float: "                    input <- getLine                    parseTest float input

*Main> mainEnter a float: 2.32.3Enter a float: 11.0Enter a float: -1-1.0Enter a float: +6.986.98

• Plus = char ' + ' *> number
  P1 *> P2 matches p1 and P2 in turn, but P1 is discarded, returning only P2.
• minus = (:) <$> char '-' <*> number
  (:) <$> p1 <*> p2 match p1 and P2 in turn, and then bind the results of two matches.
  Here P1 matches the result is the character, P2 matches the result is the string, therefore uses (:) To connect.
• float = Fmap Rd $ (+ +) <$> integer <*> decimal
  (+ +) <$> p1 <*> p2 match p1 and P2 in turn, and then bind the results two times.
  The results of the two matches here are all strings, so use (+ +) to connect.
• decimal = Option "" $ (:) <$> char '. ' <*> number
  Option "" P tries to match p if unsuccessful, returns an empty string "".

Reference links

Parsing floats with parsec

Haskell Language Learning Note (parsec) (3)