Control.Monad.Writer.Lazy

listens f m is an action that executes the action m and adds the result of applying f to the output to the value of the computation.

• listens f m = liftM (id *** f) (listen m)

censor f m is an action that executes the action m and applies the function f to its output, leaving the return value unchanged.

• censor f m = pass (liftM (\ x -> (x,f)) m)

A writer monad parameterized by the type w of output to accumulate.

The return function produces the output mempty, while m >>= k combines the outputs of the subcomputations using mappend (also known as <>):

image: images/bind-WriterT.svg

Unwrap a writer computation as a (result, output) pair. (The inverse of writer.)

Extract the output from a writer computation.

• execWriter m = snd (runWriter m)

Map both the return value and output of a computation using the given function.

• runWriter (mapWriter f m) = f (runWriter m)

A writer monad parameterized by:

• w - the output to accumulate.

• m - The inner monad.

The return function produces the output mempty, while m >>= k combines the outputs of the subcomputations using mappend (also known as <>):

image: images/bind-WriterT.svg

Extract the output from a writer computation.

• execWriterT m = liftM snd (runWriterT m)

Map both the return value and output of a computation using the given function.

• runWriterT (mapWriterT f m) = f (runWriterT m)

The Monad class defines the basic operations over a monad, a concept from a branch of mathematics known as category theory. From the perspective of a Haskell programmer, however, it is best to think of a monad as an abstract datatype of actions. Haskell's do expressions provide a convenient syntax for writing monadic expressions.

Instances of Monad should satisfy the following:

Left identity

return a >>= k = k a

Right identity

m >>= return = m

Associativity

m >>= (\x -> k x >>= h) = (m >>= k) >>= h

Furthermore, the Monad and Applicative operations should relate as follows:

• pure = return

• m1 <*> m2 = m1 >>= (\x1 -> m2 >>= (\x2 -> return (x1 x2)))

The above laws imply:

• fmap f xs  =  xs >>= return . f

• (>>) = (*>)

and that pure and (<*>) satisfy the applicative functor laws.

The instances of Monad for GHC.List.List, Maybe and System.IO.IO defined in the Prelude satisfy these laws.

Same as >>=, but with the arguments interchanged.

as >>= f == f =<< as

Map each element of a structure to a monadic action, evaluate these actions from left to right, and ignore the results. For a version that doesn't ignore the results see Data.Traversable.mapM.

mapM_ is just like traverse_, but specialised to monadic actions.

Evaluate each monadic action in the structure from left to right, and ignore the results. For a version that doesn't ignore the results see Data.Traversable.sequence.

sequence_ is just like sequenceA_, but specialised to monadic actions.

A type f is a Functor if it provides a function fmap which, given any types a and b lets you apply any function from (a -> b) to turn an f a into an f b, preserving the structure of f. Furthermore f needs to adhere to the following:

Identity

fmap id == id

Composition

fmap (f . g) == fmap f . fmap g

Note, that the second law follows from the free theorem of the type fmap and the first law, so you need only check that the former condition holds. See these articles by School of Haskell or David Luposchainsky for an explanation.

Map each element of a structure to a monadic action, evaluate these actions from left to right, and collect the results. For a version that ignores the results see Data.Foldable.mapM_.

Examples

mapM is literally a traverse with a type signature restricted to Monad. Its implementation may be more efficient due to additional power of Monad.

Evaluate each monadic action in the structure from left to right, and collect the results. For a version that ignores the results see Data.Foldable.sequence_.

Examples

Basic usage:

The first two examples are instances where the input and and output of sequence are isomorphic.

>>> sequence $ Right [1,2,3,4][Right 1,Right 2,Right 3,Right 4]

>>> sequence $ [Right 1,Right 2,Right 3,Right 4]Right [1,2,3,4]

The following examples demonstrate short circuit behavior for sequence.

>>> sequence $ Left [1,2,3,4]Left [1,2,3,4]

>>> sequence $ [Left 0, Right 1,Right 2,Right 3,Right 4]Left 0

The join function is the conventional monad join operator. It is used to remove one level of monadic structure, projecting its bound argument into the outer level.

'join bss' can be understood as the do expression

do bs <- bss
   bs

Examples

>>> join [[1, 2, 3], [4, 5, 6], [7, 8, 9]][1,2,3,4,5,6,7,8,9]

>>> join (Just (Just 3))Just 3

A common use of join is to run an IO computation returned from an GHC.Conc.STM transaction, since GHC.Conc.STM transactions can't perform IO directly. Recall that

GHC.Internal.Conc.atomically :: STM a -> IO a

is used to run GHC.Conc.STM transactions atomically. So, by specializing the types of GHC.Internal.Conc.atomically and join to

GHC.Internal.Conc.atomically :: STM (IO b) -> IO (IO b)
join       :: IO (IO b)  -> IO b

we can compose them as

join . GHC.Internal.Conc.atomically :: STM (IO b) -> IO b

to run an GHC.Conc.STM transaction and the IO action it returns.

Monads that also support choice and failure.

In many situations, the liftM operations can be replaced by uses of ap, which promotes function application.

return f `ap` x1 `ap` ... `ap` xn

is equivalent to

liftM<n> f x1 x2 ... xn

Examples

>>> pure (\x y z -> x + y * z) `ap` Just 1 `ap` Just 5 `ap` Just 10Just 51

Promote a function to a monad. This is equivalent to fmap but specialised to Monads.

Promote a function to a monad, scanning the monadic arguments from left to right.

Examples

>>> liftM2 (+) [0,1] [0,2][0,2,1,3]

>>> liftM2 (+) (Just 1) NothingNothing

>>> liftM2 (+) (+ 3) (* 2) 518

Promote a function to a monad, scanning the monadic arguments from left to right (cf. liftM2).

Promote a function to a monad, scanning the monadic arguments from left to right (cf. liftM2).

Promote a function to a monad, scanning the monadic arguments from left to right (cf. liftM2).

Conditional execution of Applicative expressions. For example,

Examples

when debug (putStrLn "Debugging")

will output the string Debugging if the Boolean value debug is True, and otherwise do nothing.

>>> putStr "pi:" >> when False (print 3.14159)pi:

When a value is bound in do-notation, the pattern on the left hand side of <- might not match. In this case, this class provides a function to recover.

A Monad without a MonadFail instance may only be used in conjunction with pattern that always match, such as newtypes, tuples, data types with only a single data constructor, and irrefutable patterns (~pat).

Instances of MonadFail should satisfy the following law: fail s should be a left zero for >>=,

fail s >>= f  =  fail s

If your Monad is also MonadPlus, a popular definition is

fail _ = mzero

fail s should be an action that runs in the monad itself, not an exception (except in instances of MonadIO). In particular, fail should not be implemented in terms of error.

void value discards or ignores the result of evaluation, such as the return value of an System.IO.IO action.

Examples

Replace the contents of a Maybe Int with unit:

>>> void NothingNothing

>>> void (Just 3)Just ()

Replace the contents of an Either Int Int with unit, resulting in an Either Int ():

>>> void (Left 8675309)Left 8675309

>>> void (Right 8675309)Right ()

Replace every element of a list with unit:

>>> void [1,2,3][(),(),()]

Replace the second element of a pair with unit:

>>> void (1,2)(1,())

Discard the result of an System.IO.IO action:

>>> mapM print [1,2]12[(),()]

>>> void $ mapM print [1,2]12

Strict version of Data.Functor.<$>.

Right-to-left composition of Kleisli arrows. (>=>), with the arguments flipped.

Note how this operator resembles function composition (.):

(.)   ::            (b ->   c) -> (a ->   b) -> a ->   c
(<=<) :: Monad m => (b -> m c) -> (a -> m b) -> a -> m c

Left-to-right composition of Kleisli arrows.

'(bs >=> cs) a' can be understood as the do expression

do b <- bs a
   cs b

or in terms of (>>=) as

bs a >>= cs

This generalizes the list-based filter function.

runIdentity (filterM (Identity . p) xs) == filter p xs

Examples

>>> filterM (\x -> do      putStrLn ("Keep: " ++ show x ++ "?")      answer <- getLine      pure (answer == "y"))    [1, 2, 3]Keep: 1?yKeep: 2?nKeep: 3?y[1,3]

>>> filterM (\x -> do      putStr (show x)      x' <- readLn      pure (x == x'))    [1, 2, 3]122233[2,3]

The foldM function is analogous to foldl, except that its result is encapsulated in a monad. Note that foldM works from left-to-right over the list arguments. This could be an issue where (>>) and the `folded function' are not commutative.

foldM f a1 [x1, x2, ..., xm]

==

do
  a2 <- f a1 x1
  a3 <- f a2 x2
  ...
  f am xm

If right-to-left evaluation is required, the input list should be reversed.

Note: foldM is the same as foldlM

Like foldM, but discards the result.

Repeat an action indefinitely.

Examples

A common use of forever is to process input from network sockets, System.IO.Handles, and channels (e.g. Control.Concurrent.MVar.MVar and Chan).

For example, here is how we might implement an echo server, using forever both to listen for client connections on a network socket and to echo client input on client connection handles:

echoServer :: Socket -> IO ()
echoServer socket = forever $ do
  client <- accept socket
  forkFinally (echo client) (\_ -> hClose client)
  where
    echo :: Handle -> IO ()
    echo client = forever $
      hGetLine client >>= hPutStrLn client

Note that "forever" isn't necessarily non-terminating. If the action is in a MonadPlus and short-circuits after some number of iterations. then forever actually returns mzero, effectively short-circuiting its caller.

The mapAndUnzipM function maps its first argument over a list, returning the result as a pair of lists. This function is mainly used with complicated data structures or a state monad.

Direct MonadPlus equivalent of filter.

Examples

The filter function is just mfilter specialized to the list monad:

filter = ( mfilter :: (a -> Bool) -> [a] -> [a] )

An example using mfilter with the Maybe monad:

>>> mfilter odd (Just 1)Just 1>>> mfilter odd (Just 2)Nothing

replicateM n act performs the action act n times, and then returns the list of results.

replicateM n (pure x) == replicate n x

Examples

>>> replicateM 3 getLinehiheyahiya["hi","heya","hiya"]

>>> import Control.Monad.State>>> runState (replicateM 3 $ state $ \s -> (s, s + 1)) 1([1,2,3],4)

Like replicateM, but discards the result.

Examples

>>> replicateM_ 3 (putStr "a")aaa

The reverse of when.

Examples

>>> do x <- getLine       unless (x == "hi") (putStrLn "hi!")comingupwithexamplesisdifficulthi!

>>> unless (pi > exp 1) NothingJust ()

The zipWithM function generalizes zipWith to arbitrary applicative functors.

zipWithM_ is the extension of zipWithM which ignores the final result.

forM_ is mapM_ with its arguments flipped. For a version that doesn't ignore the results see Data.Traversable.forM.

forM_ is just like for_, but specialised to monadic actions.

The sum of a collection of actions using (<|>), generalizing concat.

msum is just like asum, but specialised to MonadPlus.

Examples

Basic usage, using the MonadPlus instance for Maybe:

>>> msum [Just "Hello", Nothing, Just "World"]Just "Hello"

forM is mapM with its arguments flipped. For a version that ignores the results see Data.Foldable.forM_.

Conditional failure of Alternative computations. Defined by

guard True  = pure ()
guard False = empty

Examples

Common uses of guard include conditionally signalling an error in an error monad and conditionally rejecting the current choice in an Alternative-based parser.

As an example of signalling an error in the error monad Maybe, consider a safe division function safeDiv x y that returns Nothing when the denominator y is zero and Just (x `div` y) otherwise. For example:

>>> safeDiv 4 0Nothing

>>> safeDiv 4 2Just 2

A definition of safeDiv using guards, but not guard:

safeDiv :: Int -> Int -> Maybe Int
safeDiv x y | y /= 0    = Just (x `div` y)
            | otherwise = Nothing

A definition of safeDiv using guard and Monad do-notation:

safeDiv :: Int -> Int -> Maybe Int
safeDiv x y = do
  guard (y /= 0)
  return (x `div` y)

Monads having fixed points with a 'knot-tying' semantics. Instances of MonadFix should satisfy the following laws:

Purity

mfix (return . h) = return (fix h)

Left shrinking (or Tightening)

mfix (\x -> a >>= \y -> f x y) = a >>= \y -> mfix (\x -> f x y)

Sliding

mfix (liftM h . f) = liftM h (mfix (f . h))

, for strict

h

.

Nesting

mfix (\x -> mfix (\y -> f x y)) = mfix (\x -> f x x)

This class is used in the translation of the recursive do notation supported by GHC and Hugs.

fix f is the least fixed point of the function f, i.e. the least defined x such that f x = x.

When f is strict, this means that because, by the definition of strictness, f ⊥ = ⊥ and such the least defined fixed point of any strict function is ⊥.

Examples

We can write the factorial function using direct recursion as

>>> let fac n = if n <= 1 then 1 else n * fac (n-1) in fac 5120

This uses the fact that Haskell’s let introduces recursive bindings. We can rewrite this definition using fix,

Instead of making a recursive call, we introduce a dummy parameter rec; when used within fix, this parameter then refers to fix’s argument, hence the recursion is reintroduced.

>>> fix (\rec n -> if n <= 1 then 1 else n * rec (n-1)) 5120

Using fix, we can implement versions of repeat as fix . (:) and cycle as fix . (++)

>>> take 10 $ fix (0:)[0,0,0,0,0,0,0,0,0,0]

>>> map (fix (\rec n -> if n < 2 then n else rec (n - 1) + rec (n - 2))) [1..10][1,1,2,3,5,8,13,21,34,55]

Implementation Details

The current implementation of fix uses structural sharing

fix f = let x = f x in x

A more straightforward but non-sharing version would look like

fix f = f (fix f)

The class of monoids (types with an associative binary operation that has an identity). Instances should satisfy the following:

Right identity

x <> mempty = x

Left identity

mempty <> x = x

Associativity

x <> (y <> z) = (x <> y) <> z

(

Semigroup

law)

Concatenation

mconcat = foldr (<>) mempty

You can alternatively define mconcat instead of mempty, in which case the laws are:

Unit

mconcat (pure x) = x

Multiplication

mconcat (join xss) = mconcat (fmap mconcat xss)

Subclass

mconcat (toList xs) = sconcat xs

The method names refer to the monoid of lists under concatenation, but there are many other instances.

Some types can be viewed as a monoid in more than one way, e.g. both addition and multiplication on numbers. In such cases we often define newtypes and make those instances of Monoid, e.g. Data.Semigroup.Sum and Data.Semigroup.Product.

NOTE: Semigroup is a superclass of Monoid since base-4.11.0.0.

An associative operation.

Examples

>>> [1,2,3] <> [4,5,6][1,2,3,4,5,6]

>>> Just [1, 2, 3] <> Just [4, 5, 6]Just [1,2,3,4,5,6]

>>> putStr "Hello, " <> putStrLn "World!"Hello, World!

Boolean monoid under disjunction (||).

Any x <> Any y = Any (x || y)

Examples

>>> Any True <> mempty <> Any FalseAny {getAny = True}

>>> mconcat (map (\x -> Any (even x)) [2,4,6,7,8])Any {getAny = True}

>>> Any False <> memptyAny {getAny = False}

Monoid under multiplication.

Product x <> Product y == Product (x * y)

Examples

>>> Product 3 <> Product 4 <> memptyProduct {getProduct = 12}

>>> mconcat [ Product n | n <- [2 .. 10]]Product {getProduct = 3628800}

The dual of a Monoid, obtained by swapping the arguments of (<>).

Dual a <> Dual b == Dual (b <> a)

Examples

>>> Dual "Hello" <> Dual "World"Dual {getDual = "WorldHello"}

>>> Dual (Dual "Hello") <> Dual (Dual "World")Dual {getDual = Dual {getDual = "HelloWorld"}}

The monoid of endomorphisms under composition.

Endo f <> Endo g == Endo (f . g)

Examples

>>> let computation = Endo ("Hello, " ++) <> Endo (++ "!")>>> appEndo computation "Haskell""Hello, Haskell!"

>>> let computation = Endo (*3) <> Endo (+1)>>> appEndo computation 16

Monoid under addition.

Sum a <> Sum b = Sum (a + b)

Examples

>>> Sum 1 <> Sum 2 <> memptySum {getSum = 3}

>>> mconcat [ Sum n | n <- [3 .. 9]]Sum {getSum = 42}

Maybe monoid returning the rightmost non-Nothing value.

Last a is isomorphic to Dual (First a), and thus to Dual (Alt Maybe a)

Data.Semigroup.Last. The former returns the last non-Nothing, so x <> Data.Monoid.Last Nothing = x. The latter simply returns the last value, thus x <> Data.Semigroup.Last Nothing = Data.Semigroup.Last Nothing.

Examples

>>> Last (Just "hello") <> Last Nothing <> Last (Just "world")Last {getLast = Just "world"}

>>> Last Nothing <> memptyLast {getLast = Nothing}

Maybe monoid returning the leftmost non-Nothing value.

First a is isomorphic to Alt Maybe a, but precedes it historically.

Beware that Data.Monoid.First is different from Data.Semigroup.First. The former returns the first non-Nothing, so Data.Monoid.First Nothing <> x = x. The latter simply returns the first value, thus Data.Semigroup.First Nothing <> x = Data.Semigroup.First Nothing.

Examples

>>> First (Just "hello") <> First Nothing <> First (Just "world")First {getFirst = Just "hello"}

>>> First Nothing <> memptyFirst {getFirst = Nothing}

Boolean monoid under conjunction (&&).

All x <> All y = All (x && y)

Examples

>>> All True <> mempty <> All False)All {getAll = False}

>>> mconcat (map (\x -> All (even x)) [2,4,6,7,8])All {getAll = False}

>>> All True <> memptyAll {getAll = True}

This data type witnesses the lifting of a Monoid into an Applicative pointwise.

Examples

>>> Ap (Just [1, 2, 3]) <> Ap NothingAp {getAp = Nothing}

>>> Ap [Sum 10, Sum 20] <> Ap [Sum 1, Sum 2]Ap {getAp = [Sum {getSum = 11},Sum {getSum = 12},Sum {getSum = 21},Sum {getSum = 22}]}

Monoid under <|>.

Alt l <> Alt r == Alt (l <|> r)

Examples

>>> Alt (Just 12) <> Alt (Just 24)Alt {getAlt = Just 12}

>>> Alt Nothing <> Alt (Just 24)Alt {getAlt = Just 24}