diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -1,5 +1,6 @@ # hw-prim -[![CircleCI](https://circleci.com/gh/haskell-works/hw-prim/tree/0-branch.svg?style=svg)](https://circleci.com/gh/haskell-works/hw-prim/tree/0-branch) +[![CircleCI](https://circleci.com/gh/haskell-works/hw-prim/tree/master.svg?style=svg)](https://circleci.com/gh/haskell-works/hw-prim/tree/master) +[![Travis](https://travis-ci.org/haskell-works/hw-prim.svg?branch=master)](https://travis-ci.org/haskell-works/hw-prim) Primitive types library. diff --git a/hw-prim.cabal b/hw-prim.cabal --- a/hw-prim.cabal +++ b/hw-prim.cabal @@ -1,13 +1,11 @@ --- This file has been generated from package.yaml by hpack version 0.20.0. +-- This file has been generated from package.yaml by hpack version 0.18.1. -- -- see: https://github.com/sol/hpack --- --- hash: 3da5a5c4ec15aafba41e3a027d8092886e0bca1ad4a9bcb85767e43bf4684b59 name: hw-prim -version: 0.6.2.0 +version: 0.6.2.2 synopsis: Primitive functions and data types -description: Please see README.md +description: Primitive functions and data types. category: Data stability: Experimental homepage: http://github.com/haskell-works/hw-prim#readme @@ -17,6 +15,7 @@ copyright: 2016-2018 John Ky license: BSD3 license-file: LICENSE +tested-with: GHC == 8.4.2, GHC == 8.2.2, GHC == 8.0.2, GHC == 7.10.3 build-type: Simple cabal-version: >= 1.10 @@ -35,12 +34,14 @@ library hs-source-dirs: src + other-extensions: AllowAmbiguousTypes ghc-options: -Wall -O2 -msse4.2 build-depends: - base >=4 && <5 - , bytestring - , mmap - , vector + base >=4.8 && <5 + , bytestring >= 0.9 && < 0.11 + , mmap >= 0.5 && < 0.6 + , vector >= 0.12 && < 0.13 + , transformers >= 0.4 && < 0.6 if flag(bounds-checking-enabled) cpp-options: -DBOUNDS_CHECKING_ENABLED exposed-modules: @@ -73,6 +74,7 @@ HaskellWorks.Data.Vector.BoxedVectorLike HaskellWorks.Data.Vector.StorableVectorLike other-modules: + HaskellWorks.Data.ByteString.Lazy HaskellWorks.Data.Char Paths_hw_prim default-language: Haskell2010 @@ -82,19 +84,21 @@ main-is: Spec.hs hs-source-dirs: test + other-extensions: AllowAmbiguousTypes ghc-options: -Wall -threaded -rtsopts -with-rtsopts=-N build-depends: - QuickCheck - , base >=4 && <5 - , bytestring - , directory - , exceptions - , hedgehog - , hspec - , hw-hspec-hedgehog + base >=4.8 && <5 + , bytestring >= 0.9 && < 0.11 + , mmap >= 0.5 && < 0.6 + , vector >= 0.12 && < 0.13 + , transformers >= 0.4 && < 0.6 + , QuickCheck >= 2.10 && < 2.12 + , directory >= 1.2 && < 1.4 + , exceptions >= 0.8 && < 0.11 + , hedgehog >= 0.5 && < 0.7 + , hspec >= 2.4 && <2.6 + , hw-hspec-hedgehog >= 0.1 && < 0.2 , hw-prim - , mmap - , vector if flag(bounds-checking-enabled) cpp-options: -DBOUNDS_CHECKING_ENABLED other-modules: @@ -102,7 +106,6 @@ HaskellWorks.Data.FromForeignRegionSpec HaskellWorks.Data.SearchSpec HaskellWorks.Data.Vector.AsVector64sSpec - Paths_hw_prim default-language: Haskell2010 benchmark bench @@ -110,16 +113,17 @@ main-is: Main.hs hs-source-dirs: bench + other-extensions: AllowAmbiguousTypes ghc-options: -Wall -O2 -msse4.2 build-depends: - base >=4 && <5 - , bytestring - , criterion + base >=4.8 && <5 + , bytestring >= 0.9 && < 0.11 + , mmap >= 0.5 && < 0.6 + , vector >= 0.12 && < 0.13 + , transformers >= 0.4 && < 0.6 + , criterion >= 1.2 && < 1.5 , hw-prim - , mmap - , vector + , mmap >= 0.5 && < 0.6 if flag(bounds-checking-enabled) cpp-options: -DBOUNDS_CHECKING_ENABLED - other-modules: - Paths_hw_prim default-language: Haskell2010 diff --git a/src/HaskellWorks/Data/ByteString.hs b/src/HaskellWorks/Data/ByteString.hs --- a/src/HaskellWorks/Data/ByteString.hs +++ b/src/HaskellWorks/Data/ByteString.hs @@ -1,8 +1,41 @@ +{-# LANGUAGE FlexibleInstances #-} + module HaskellWorks.Data.ByteString ( chunkedBy + , ToByteString(..) ) where -import qualified Data.ByteString as BS +import Data.Word + +import qualified Data.ByteString as BS +import qualified Data.ByteString.Internal as BSI +import qualified Data.Vector.Storable as DVS + +class ToByteString a where + toByteString :: a -> BS.ByteString + +instance ToByteString BS.ByteString where + toByteString = id + +instance ToByteString (DVS.Vector Word8) where + toByteString v = case DVS.unsafeToForeignPtr v of + (fptr, start, offset) -> BSI.fromForeignPtr fptr start offset + {-# INLINE toByteString #-} + +instance ToByteString (DVS.Vector Word16) where + toByteString v = case DVS.unsafeToForeignPtr (DVS.unsafeCast v :: DVS.Vector Word8) of + (fptr, start, offset) -> BSI.fromForeignPtr fptr start offset + {-# INLINE toByteString #-} + +instance ToByteString (DVS.Vector Word32) where + toByteString v = case DVS.unsafeToForeignPtr (DVS.unsafeCast v :: DVS.Vector Word8) of + (fptr, start, offset) -> BSI.fromForeignPtr fptr start offset + {-# INLINE toByteString #-} + +instance ToByteString (DVS.Vector Word64) where + toByteString v = case DVS.unsafeToForeignPtr (DVS.unsafeCast v :: DVS.Vector Word8) of + (fptr, start, offset) -> BSI.fromForeignPtr fptr start offset + {-# INLINE toByteString #-} -- | Chunk a @bs into list of smaller byte strings of no more than @n elements chunkedBy :: Int -> BS.ByteString -> [BS.ByteString] diff --git a/src/HaskellWorks/Data/ByteString/Lazy.hs b/src/HaskellWorks/Data/ByteString/Lazy.hs new file mode 100644 --- /dev/null +++ b/src/HaskellWorks/Data/ByteString/Lazy.hs @@ -0,0 +1,48 @@ +{-# LANGUAGE FlexibleInstances #-} + +module HaskellWorks.Data.ByteString.Lazy where + +import Data.Word +import HaskellWorks.Data.ByteString (ToByteString (..)) + +import qualified Data.ByteString.Lazy as LBS +import qualified Data.Vector.Storable as DVS + +class ToLazyByteString a where + toLazyByteString :: a -> LBS.ByteString + +instance ToLazyByteString LBS.ByteString where + toLazyByteString = id + {-# INLINE toLazyByteString #-} + +instance ToLazyByteString (DVS.Vector Word8) where + toLazyByteString = LBS.fromStrict . toByteString + {-# INLINE toLazyByteString #-} + +instance ToLazyByteString (DVS.Vector Word16) where + toLazyByteString = LBS.fromStrict . toByteString + {-# INLINE toLazyByteString #-} + +instance ToLazyByteString (DVS.Vector Word32) where + toLazyByteString = LBS.fromStrict . toByteString + {-# INLINE toLazyByteString #-} + +instance ToLazyByteString (DVS.Vector Word64) where + toLazyByteString = LBS.fromStrict . toByteString + {-# INLINE toLazyByteString #-} + +instance ToLazyByteString [DVS.Vector Word8] where + toLazyByteString vs = LBS.fromChunks (toByteString <$> vs) + {-# INLINE toLazyByteString #-} + +instance ToLazyByteString [DVS.Vector Word16] where + toLazyByteString vs = LBS.fromChunks (toByteString <$> vs) + {-# INLINE toLazyByteString #-} + +instance ToLazyByteString [DVS.Vector Word32] where + toLazyByteString vs = LBS.fromChunks (toByteString <$> vs) + {-# INLINE toLazyByteString #-} + +instance ToLazyByteString [DVS.Vector Word64] where + toLazyByteString vs = LBS.fromChunks (toByteString <$> vs) + {-# INLINE toLazyByteString #-} diff --git a/src/HaskellWorks/Data/Vector/AsVector64s.hs b/src/HaskellWorks/Data/Vector/AsVector64s.hs --- a/src/HaskellWorks/Data/Vector/AsVector64s.hs +++ b/src/HaskellWorks/Data/Vector/AsVector64s.hs @@ -8,6 +8,7 @@ ( AsVector64s(..) ) where +import Control.Applicative ((<$>)) import Control.Monad.ST import Data.Word import Foreign.ForeignPtr @@ -32,7 +33,7 @@ {-# INLINE asVector64s #-} bytestringsToVectors :: Int -> [BS.ByteString] -> [DVS.Vector Word64] -bytestringsToVectors n bss = go bss +bytestringsToVectors n = go where go :: [BS.ByteString] -> [DVS.Vector Word64] go bs = case DVS.createT (buildOneVector n bs) of (cs, ws) -> if DVS.length ws > 0 @@ -51,16 +52,15 @@ where go :: [BS.ByteString] -> DVSM.MVector s Word8 -> ST s [BS.ByteString] go ts v = if DVSM.length v > 0 then case ts of - (u:us) -> do - if BS.length u <= DVSM.length v - then case DVSM.splitAt (BS.length u) v of - (va, vb) -> do - DVSM.copy va (byteStringToVector8 u) - go us vb - else case BS.splitAt (DVSM.length v) u of - (ua, ub) -> do - DVSM.copy v (byteStringToVector8 ua) - return (ub:us) + (u:us) -> if BS.length u <= DVSM.length v + then case DVSM.splitAt (BS.length u) v of + (va, vb) -> do + DVSM.copy va (byteStringToVector8 u) + go us vb + else case BS.splitAt (DVSM.length v) u of + (ua, ub) -> do + DVSM.copy v (byteStringToVector8 ua) + return (ub:us) [] -> do DVSM.set v 0 return []