Control.Concurrent.SCC.Foundation

Module Foundation defines the pipe computations and their basic building blocks.

Classes

class Monad m => ParallelizableMonad m where

Methods

parallelize :: m a -> m b -> m (a, b)

Instances

ParallelizableMonad IO

ParallelizableMonad Maybe

ParallelizableMonad Identity

ParallelizableMonad m => ParallelizableMonad (Pipe context m)

Types

data Pipe context m r

Pipe represents the type of monadic computations that can be split into co-routining computations using function pipe. The context type parameter delimits the scope of the computation.

Instances

Monad m => Monad (Pipe context m)

ParallelizableMonad m => ParallelizableMonad (Pipe context m)

data Source context x

A Source is the read-only end of a Pipe communication channel.

data Sink context x

A Sink is the write-only end of a Pipe communication channel.

Flow-control functions

pipe :: forall context x m r1 r2. Monad m => Producer context m x r1 -> Consumer context m x r2 -> Pipe context m (r1, r2)

The pipe function splits the computation into two concurrent parts, producer and consumer. The producer is given a Sink to put values into, and consumer a Source to get those values from. Once producer and consumer both complete, pipe returns their paired results.

pipeD :: forall c x m r1 r2. Monad m => String -> Producer c m x r1 -> Consumer c m x r2 -> Pipe c m (r1, r2)

The pipeD function is same as pipe, with an additional description argument.

pipeP :: forall c x m r1 r2. ParallelizableMonad m => Producer c m x r1 -> Consumer c m x r2 -> Pipe c m (r1, r2)

The pipeP function is equivalent to pipe, except the producer and consumer are run in parallel if resources allow.

get :: forall context x m r. (Monad m, Typeable x) => Source context x -> Pipe context m (Maybe x)

Function get tries to get a value from the given Source argument. The intervening Pipe computations suspend all the way to the pipe function invocation that created the source. The result of get is Nothing iff the argument source is empty.

getSuccess

Arguments

:: forall context x m . (Monad m, Typeable x)

=> Source context x

-> (x -> Pipe context m ())

Success continuation

-> Pipe context m ()

get' :: forall context x m r. (Monad m, Typeable x) => Source context x -> Pipe context m x

Function get' assumes that the argument source is not empty and returns the value the source yields. If the source is empty, the function throws an error.

canPut :: forall context x m r. (Monad m, Typeable x) => Sink context x -> Pipe context m Bool

Function canPut checks if the argument sink accepts values, i.e., whether a put operation would succeed on the sink.

put :: forall context x m r. (Monad m, Typeable x) => Sink context x -> x -> Pipe context m Bool

Function put tries to put a value into the given sink. The intervening Pipe computations suspend up to the pipe invocation that has created the argument sink. The result of put indicates whether the operation succeded.

liftPipe :: forall context m r. Monad m => m r -> Pipe context m r

Function liftPipe lifts a value of the underlying monad type into a Pipe computation.

runPipes :: forall m r. Monad m => (forall context. Pipe context m r) -> m r

Function runPipes runs the given computation involving pipes and returns the final result. The context argument ensures that no suspended computation can escape its scope.

Utility functions

cond :: a -> a -> Bool -> a

A utility function wrapping if-then-else, useful for handling monadic truth values

whenNull :: forall a m. Monad m => m [a] -> [a] -> m [a]

A utility function, useful for handling monadic list values where empty list means success

pour :: forall c x m. (Monad m, Typeable x) => Source c x -> Sink c x -> Pipe c m ()

pour copies all data from the source argument into the sink argument, as long as there is anything to copy and the sink accepts it.

pourMap :: forall c x y m. (Monad m, Typeable x, Typeable y) => (x -> y) -> Source c x -> Sink c y -> Pipe c m ()

pourMap is like pour that applies the function f to each argument before passing it into the sink.

pourMapMaybe :: forall c x y m. (Monad m, Typeable x, Typeable y) => (x -> Maybe y) -> Source c x -> Sink c y -> Pipe c m ()

pourMapMaybe is to pourMap like Data.Maybe.mapMaybe is to Data.List.Map.

tee :: (Monad m, Typeable x) => Source c x -> Sink c x -> Sink c x -> Pipe c m ()

tee is similar to pour except it distributes every input value from the source arguments into both sink1 and sink2.

getList :: forall x c m. (Monad m, Typeable x) => Source c x -> Pipe c m [x]

getList returns the list of all values generated by the source.

putList :: forall x c m. (Monad m, Typeable x) => [x] -> Sink c x -> Pipe c m [x]

putList puts entire list into its sink argument, as long as the sink accepts it. The remainder that wasn't accepted by the sink is the result value.

putQueue :: forall c m x. (Monad m, Typeable x) => Seq x -> Sink c x -> Pipe c m [x]

Like putList, except it puts the contents of the given Seq into the sink.

consumeAndSuppress :: forall x c m. (Monad m, Typeable x) => Source c x -> Pipe c m ()

consumeAndSuppress consumes the entire source ignoring the values it generates.