Copyright	(c) 2019 Composewell Technologies
License	BSD3
Maintainer	streamly@composewell.com
Stability	experimental
Portability	GHC
Safe Haskell	None
Language	Haskell2010

Streamly.Internal.Data.IORef

Description

A mutable variable in a mutation capable monad (IO) holding a Unboxed value. This allows fast modification because of unboxed storage.

Multithread Consistency Notes

In general, any value that straddles a machine word cannot be guaranteed to be consistently read from another thread without a lock. GHC heap objects are always machine word aligned, therefore, a IORef is also word aligned. On a 64-bit platform, writing a 64-bit aligned type from one thread and reading it from another thread should give consistent old or new value. The same holds true for 32-bit values on a 32-bit platform.

Synopsis

data IORef a
newIORef :: Unbox a => a -> IO (IORef a)
writeIORef :: Unbox a => IORef a -> a -> IO ()
modifyIORef' :: Unbox a => IORef a -> (a -> a) -> IO ()
readIORef :: Unbox a => IORef a -> IO a
pollGenericIORef :: forall (m :: Type -> Type) a. (MonadIO m, Unbox a) => IORef a -> Stream m a
pollIORefInt :: forall (m :: Type -> Type). MonadIO m => IORef Int -> Stream m Int

Documentation

data IORef a Source #

An IORef holds a single Unbox-able value.

newIORef :: Unbox a => a -> IO (IORef a) Source #

Create a new IORef.

Pre-release

writeIORef :: Unbox a => IORef a -> a -> IO () Source #

Write a value to an IORef.

Pre-release

modifyIORef' :: Unbox a => IORef a -> (a -> a) -> IO () Source #

Modify the value of an IORef using a function with strict application.

Pre-release

readIORef :: Unbox a => IORef a -> IO a Source #

Read a value from an IORef.

Pre-release

pollGenericIORef :: forall (m :: Type -> Type) a. (MonadIO m, Unbox a) => IORef a -> Stream m a Source #

Internal, do not use.

pollIORefInt :: forall (m :: Type -> Type). MonadIO m => IORef Int -> Stream m Int Source #

Generate a stream by continuously reading the IORef.

This operation reads the IORef without any synchronization. It can be assumed to be atomic because the size fits into machine register size. We are assuming that compiler uses single instructions to access the memory. It may read stale values though until caches are synchronised in a multiprocessor architecture.

Pre-release