Movatterモバイル変換

{-# LANGUAGE Trustworthy #-}{-# LANGUAGE MagicHash #-}------------------------------------------------------------------------------- |-- Module      :  Data.Unique-- Copyright   :  (c) The University of Glasgow 2001-- License     :  BSD-style (see the file libraries/base/LICENSE)---- Maintainer  :  libraries@haskell.org-- Stability   :  experimental-- Portability :  non-portable---- An abstract interface to a unique symbol generator.-------------------------------------------------------------------------------moduleData.Unique(-- * Unique objectsUnique,newUnique,hashUnique)whereimportSystem.IO.Unsafe(unsafePerformIO)importGHC.BaseimportGHC.NumimportData.IORef-- | An abstract unique object.  Objects of type 'Unique' may be-- compared for equality and ordering and hashed into 'Int'.---- >>> :{-- do x <- newUnique--    print (x == x)--    y <- newUnique--    print (x == y)-- :}-- True-- FalsenewtypeUnique=UniqueIntegerderiving(Eq,Ord)uniqSource::IORefIntegeruniqSource=unsafePerformIO(newIORef0){-# NOINLINEuniqSource#-}-- | Creates a new object of type 'Unique'.  The value returned will-- not compare equal to any other value of type 'Unique' returned by-- previous calls to 'newUnique'.  There is no limit on the number of-- times 'newUnique' may be called.newUnique::IOUniquenewUnique=dor<-atomicModifyIORef'uniqSource$\x->letz=x+1in(z,z)return(Uniquer)-- SDM (18/3/2010): changed from MVar to STM.  This fixes--  1. there was no async exception protection--  2. there was a space leak (now new value is strict)--  3. using atomicModifyIORef would be slightly quicker, but can--     suffer from adverse scheduling issues (see #3838)--  4. also, the STM version is faster.-- SDM (30/4/2012): changed to IORef using atomicModifyIORef.  Reasons:--  1. STM version could not be used inside unsafePerformIO, if it--     happened to be poked inside an STM transaction.--  2. IORef version can be used with unsafeIOToSTM inside STM,--     because if the transaction retries then we just get a new--     Unique.--  3. IORef version is very slightly faster.-- IGL (08/06/2013): changed to using atomicModifyIORef' instead.--  This feels a little safer, from the point of view of not leaking--  memory, but the resulting core is identical.-- | Hashes a 'Unique' into an 'Int'.  Two 'Unique's may hash to the-- same value, although in practice this is unlikely.  The 'Int'-- returned makes a good hash key.hashUnique::Unique->InthashUnique(Uniquei)=I#(hashIntegeri)