|Portability||non-portable (type families)|
MaybeT monad. See
http://www.haskell.org/haskellwiki/New_monads/MaybeT for more widely-used
Functor instance and our implementation of
borrowed from there.
- Computation type:
- Computations which may fail or return nothing.
- Binding strategy:
- Failure returns the value
Nothing, bypassing any bound functions which follow. Success returns a value wrapped in
- Useful for:
- Building computations from steps which may fail. No error
information is returned. (If error information is required, see
newtype MaybeT m a
The parameterizable maybe monad, obtained by composing an arbitrary
monad with the
Computations are actions that may produce a value or fail.
return function yields a successful computation, while
sequences two subcomputations, failing on the first error.
Here is an example that shows how to use
MaybeT to propagate an
end-of-file condition in the IO monad. In the example below, both
failIfQuit may cause a failure, which will propagate
main without further intervention.
import System.Console.Readline import Data.Maybe import Control.Monad import Control.Monad.Trans import Control.Monad.Maybe -- 'MaybeIO' is the type of computations which do IO, and which may fail. type MaybeIO = MaybeT IO -- 'readline' already has type 'String -> IO (Maybe String)'; we just need -- to wrap it. maybeReadLine :: String -> MaybeIO String maybeReadLine prompt = MaybeT (readline prompt) -- Fail if 'str' equals "quit". failIfQuit :: (Monad m) => String -> m () failIfQuit str = when (str == "quit") (fail "Quitting") -- This task may fail in several places. Try typing Control-D or "quit" at -- any prompt. concatTwoInputs :: MaybeIO () concatTwoInputs = do s1 <- maybeReadLine "String 1> " failIfQuit s1 s2 <- maybeReadLine "String 2> " failIfQuit s2 liftIO (putStrLn ("Concatenated: " ++ s1 ++ s2)) -- Loop until failure. main :: IO () main = do result <- runMaybeT concatTwoInputs if isNothing result then putStrLn "Bye!" else main