diff --git a/.travis.yml b/.travis.yml deleted file mode 100644 --- a/.travis.yml +++ /dev/null @@ -1,1 +0,0 @@ -language: haskell diff --git a/CHANGELOG.markdown b/CHANGELOG.markdown new file mode 100644 --- /dev/null +++ b/CHANGELOG.markdown @@ -0,0 +1,28 @@ +0.3.8 [2021.02.17] +------------------ +* Add an `IsString CharSet` instance. + +0.3.7.1 +------- +* Minor haddock fixup. + +0.3.7 +----- +* Switched to derived Typeable for GHC 7.8 compatibility + +0.3.6 +----- +* Removed some duplicated blocks in `Data.CharSet.Unicode.Block.blocks`, see issue #3. + +0.3.5.1 +------- +* Updated dependencies to support GHC 7.8 + +0.3.5 +----- +* Claim to be Trustworthy. + +0.3.3 +----- +* Removed upper bounds on my other packages + diff --git a/README.markdown b/README.markdown new file mode 100644 --- /dev/null +++ b/README.markdown @@ -0,0 +1,15 @@ +charsets +======== + +[![Hackage](https://img.shields.io/hackage/v/charset.svg)](https://hackage.haskell.org/package/charset) [![Build Status](https://github.com/ekmett/charset/workflows/Haskell-CI/badge.svg)](https://github.com/ekmett/charset/actions?query=workflow%3AHaskell-CI) + +Fast utf-8 character sets for Haskell represented as complemented PATRICIA tries. + +Contact Information +------------------- + +Contributions and bug reports are welcome! + +Please feel free to contact me through github or on the #haskell IRC channel on irc.freenode.net. + +-Edward Kmett diff --git a/charset.cabal b/charset.cabal --- a/charset.cabal +++ b/charset.cabal @@ -1,35 +1,50 @@ name: charset -version: 0.3.7.1 +version: 0.3.8 license: BSD3 license-File: LICENSE copyright: (c) Edward Kmett 2010-2012 author: Edward Kmett maintainer: ekmett@gmail.com -cabal-version: >= 1.6 +cabal-version: >= 1.10 stability: Experimental category: Data homepage: http://github.com/ekmett/charset bug-reports: http://github.com/ekmett/charset/issues synopsis: Fast unicode character sets based on complemented PATRICIA tries -description: Fast unicode character sets based on complemented PATRICIA tries +description: Fast unicode character sets based on complemented PATRICIA tries. build-type: Simple -extra-source-files: .travis.yml +extra-source-files: CHANGELOG.markdown, README.markdown +tested-with: + GHC ==7.0.4 + || ==7.2.2 + || ==7.4.2 + || ==7.6.3 + || ==7.8.4 + || ==7.10.3 + || ==8.0.2 + || ==8.2.2 + || ==8.4.4 + || ==8.6.5 + || ==8.8.3 + || ==8.10.1 + source-repository head type: git location: git://github.com/ekmett/charset.git library - extensions: CPP + default-extensions: CPP other-extensions: MagicHash, BangPatterns build-depends: base >= 4 && < 5, array >= 0.2 && < 0.6, - bytestring >= 0.9 && < 0.11, - containers >= 0.2 && < 0.6, - semigroups >= 0.8.3.1 && < 1, + bytestring >= 0.9 && < 0.12, + containers >= 0.2 && < 0.7, unordered-containers >= 0.1.4.6 && < 0.3 + if impl(ghc < 8.0) + build-depends: semigroups >= 0.8.3.1 && < 1 exposed-modules: Data.CharSet @@ -44,3 +59,4 @@ hs-source-dirs: src ghc-options: -Wall -fspec-constr -fdicts-cheap -O2 + default-language: Haskell2010 diff --git a/src/Data/CharSet.hs b/src/Data/CharSet.hs --- a/src/Data/CharSet.hs +++ b/src/Data/CharSet.hs @@ -13,12 +13,12 @@ -- Stability : experimental -- Portability : portable -- --- Fast set membership tests for 'Char' values --- --- Stored as a (possibly negated) IntMap and a fast set used for the head byte. +-- A CharSet is an /efficient/ representation of a set of 'Char' values +-- designed for fast membership tests. -- --- The set of valid (possibly negated) head bytes is stored unboxed as a 32-byte --- bytestring-based lookup table. +-- As an example @build isAlpha@ will create a set of alphabetic characters. +-- We can then use 'member' on the generated set to /efficiently/ test if a +-- given @Char@ represents an alphabetic character. -- -- Designed to be imported qualified: -- @@ -75,6 +75,11 @@ , toArray ) where +#if defined(__GLASGOW_HASKELL__) && __GLASGOW_HASKELL__ >= 608 +import Data.String (IsString(..)) +-- <<< -XOverloadedStrings >>> was introduced by GHC 6.8.1 +#endif + import Data.Array.Unboxed hiding (range) import Data.Data import Data.Function (on) @@ -91,9 +96,22 @@ import qualified Prelude as P import Text.Read -data CharSet = CharSet !Bool {-# UNPACK #-} !ByteSet !IntSet +-- | Stored as a (possibly negated) IntSet and a fast set used for the head byte. +-- +-- The set of valid (possibly negated) head bytes is stored unboxed as a 32-byte +-- bytestring-based lookup table. +data CharSet = CharSet + !Bool -- Whether ByteSet and IntSet are negated + !ByteSet -- Set of head bytes, unboxed + !IntSet -- Set of characters in the charset deriving Typeable +#if defined(__GLASGOW_HASKELL__) && __GLASGOW_HASKELL__ >= 608 +-- | @= CharSet.`fromList`@ +instance IsString CharSet where + fromString = fromList +#endif + charSet :: Bool -> IntSet -> CharSet charSet b s = CharSet b (ByteSet.fromList (fmap headByte (I.toAscList s))) s @@ -113,6 +131,11 @@ (\\) :: CharSet -> CharSet -> CharSet (\\) = difference +-- | Applies a predicate across the whole range of possible character values +-- to create a set of only those characters which satisfy the predicate. +-- +-- As an example @build isAlpha@ will generate a CharSet of all +-- alphabetic characters. build :: (Char -> Bool) -> CharSet build p = fromDistinctAscList $ P.filter p [minBound .. maxBound] {-# INLINE build #-} diff --git a/src/Data/CharSet/ByteSet.hs b/src/Data/CharSet/ByteSet.hs --- a/src/Data/CharSet/ByteSet.hs +++ b/src/Data/CharSet/ByteSet.hs @@ -14,9 +14,11 @@ -- Stability : experimental -- Portability : non-portable (BangPatterns, MagicHash) -- --- Fast set membership tests for byte values, The set representation is --- unboxed for efficiency and uses a lookup table. This is a fairly minimal --- API. You probably want to use CharSet. +-- Fast set membership tests for byte values. The set representation is +-- unboxed for efficiency and uses one bit per byte to represent the presence +-- or absence of a byte in the set. +-- +-- This is a fairly minimal API. You probably want to use CharSet. ----------------------------------------------------------------------------- module Data.CharSet.ByteSet ( @@ -38,7 +40,11 @@ newtype ByteSet = ByteSet B.ByteString deriving (Eq, Ord, Show) -data I = I {-# UNPACK #-} !Int {-# UNPACK #-} !Word8 +-- | Representation of the index of a bit inside a bytestring +-- in terms of a byte index and a bit index inside the byte +data I = I + {-# UNPACK #-} !Int -- byte index + {-# UNPACK #-} !Word8 -- bit index shiftR :: Int -> Int -> Int shiftR (I# x#) (I# i#) = I# (x# `iShiftRA#` i#) @@ -46,6 +52,7 @@ shiftL :: Word8 -> Int -> Word8 shiftL (W8# x#) (I# i#) = W8# (narrow8Word# (x# `shiftL#` i#)) +-- | Convert a bit index to a byte index and bit index inside the byte index :: Int -> I index i = I (i `shiftR` 3) (1 `shiftL` (i .&. 7)) {-# INLINE index #-}