diff --git a/Changelog.md b/Changelog.md --- a/Changelog.md +++ b/Changelog.md @@ -1,3 +1,10 @@ +## 0.4.0 (November 2021) + +* Bump `unicode-data` to `0.2` +* Allow `text 2.0` +* Update to Unicode version 14.0.0 +* Drop support for GHC 7.10 + ## 0.3.8 * Allow ghc-prim 0.7 diff --git a/Data/Unicode/Internal/NormalizeStream.hs b/Data/Unicode/Internal/NormalizeStream.hs --- a/Data/Unicode/Internal/NormalizeStream.hs +++ b/Data/Unicode/Internal/NormalizeStream.hs @@ -1,5 +1,6 @@ {-# OPTIONS_GHC -funbox-strict-fields #-} {-# LANGUAGE BangPatterns #-} +{-# LANGUAGE CPP #-} {-# LANGUAGE LambdaCase #-} {-# LANGUAGE TupleSections #-} -- | @@ -22,7 +23,6 @@ ) where -import Data.Bits (shiftR) import Data.Char (chr, ord) import GHC.ST (ST(..)) import GHC.Types (SPEC(..)) @@ -30,13 +30,21 @@ import qualified Data.Text.Array as A import qualified Unicode.Char as UC +#if MIN_VERSION_text(2,0,0) +import Data.Text.Internal.Fusion (stream) +#else +import Data.Bits (shiftR) +import Data.Text.Internal.Unsafe.Char (unsafeChr) +import Data.Text.Internal.Fusion.Size (betweenSize) +import Data.Text.Internal.Encoding.Utf16 (chr2) +#endif + -- Internal modules import Data.Text.Internal (Text(..)) -import Data.Text.Internal.Fusion.Size (betweenSize, upperBound) +import Data.Text.Internal.Fusion.Size (upperBound) import Data.Text.Internal.Fusion.Types (Step(..), Stream(..)) import Data.Text.Internal.Private (runText) -import Data.Text.Internal.Unsafe.Char (unsafeChr, unsafeWrite) -import Data.Text.Internal.Encoding.Utf16 (chr2) +import Data.Text.Internal.Unsafe.Char (unsafeWrite) ------------------------------------------------------------------------------- -- Reorder buffer to hold characters till the next starter boundary @@ -142,6 +150,7 @@ n <- unsafeWrite arr j c return (j + n, Empty) +#if !MIN_VERSION_text(2,0,0) -- | /O(n)/ Convert a 'Text' into a 'Stream Char'. stream :: Text -> Stream Char stream (Text arr off len) = Stream next off (betweenSize (len `shiftR` 1) len) @@ -158,6 +167,7 @@ n = A.unsafeIndex arr i n2 = A.unsafeIndex arr (i + 1) {-# INLINE [0] stream #-} +#endif -- | /O(n)/ Convert a 'Stream Char' into a decompose-normalized 'Text'. unstream :: UC.DecomposeMode -> Stream Char -> Text diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -5,7 +5,7 @@ [![Windows Build status](https://ci.appveyor.com/api/projects/status/5wov8m1m0asvbv32?svg=true)](https://ci.appveyor.com/project/harendra-kumar/unicode-transforms) [![Coverage Status](https://coveralls.io/repos/composewell/unicode-transforms/badge.svg?branch=master&service=github)](https://coveralls.io/github/composewell/unicode-transforms?branch=master) -Fast Unicode 13.0.0 normalization in Haskell (NFC, NFKC, NFD, NFKD). +Fast Unicode 14.0.0 normalization in Haskell (NFC, NFKC, NFD, NFKD). ## What is normalization? diff --git a/download-ucd-files.sh b/download-ucd-files.sh --- a/download-ucd-files.sh +++ b/download-ucd-files.sh @@ -4,11 +4,11 @@ # we used to generate them earlier are exactly the same as the ones we are # downloading. To ensure that verfication of the checksum is necessary. -VERSION=13.0.0 +VERSION=14.0.0 SRC_FILE="https://www.unicode.org/Public/$VERSION/ucd/NormalizationTest.txt" DST_FILE="test/data/NormalizationTest.txt" -SRC_FILE_CKSUM="d60ee55dd9169444652e48d337109cc814ecc59a9d3122eedddf7de388f2e01d" +SRC_FILE_CKSUM="7cb30cc2abe6c29c292b99095865d379ce1213045c78c4ff59c7e9391bbe2331" # Download the file diff --git a/test/data/NormalizationTest.txt b/test/data/NormalizationTest.txt # file too large to diff: test/data/NormalizationTest.txt diff --git a/unicode-transforms.cabal b/unicode-transforms.cabal --- a/unicode-transforms.cabal +++ b/unicode-transforms.cabal @@ -1,9 +1,9 @@ cabal-version: 2.2 name: unicode-transforms -version: 0.3.8 +version: 0.4.0 synopsis: Unicode normalization -description: Fast Unicode 13.0.0 normalization in Haskell (NFC, NFKC, NFD, NFKD). +description: Fast Unicode 14.0.0 normalization in Haskell (NFC, NFKC, NFD, NFKD). category: Data,Text,Unicode stability: Experimental homepage: http://github.com/composewell/unicode-transforms @@ -14,14 +14,14 @@ 2014–2015 Antonio Nikishaev license: BSD-3-Clause license-file: LICENSE -tested-with: GHC==7.10.3 - , GHC==8.0.2 +tested-with: GHC==8.0.2 , GHC==8.2.2 , GHC==8.4.4 , GHC==8.6.5 , GHC==8.8.4 - , GHC==8.10.3 + , GHC==8.10.7 , GHC==9.0.1 + , GHC==9.2.1 build-type: Simple extra-source-files: Changelog.md @@ -82,14 +82,14 @@ hs-source-dirs: . ghc-options: -Wall -fwarn-identities -fwarn-incomplete-record-updates -fwarn-incomplete-uni-patterns -fwarn-tabs build-depends: - base >=4.8 && <5 - , unicode-data >= 0.1 && <0.2 - , bytestring >=0.9 && <0.12 - , ghc-prim >= 0.2 && < 0.8 + base >= 4.8 && < 5 + , unicode-data >= 0.2 && < 0.3 + , bytestring >= 0.9 && < 0.12 + , ghc-prim >= 0.2 && < 0.9 -- We depend on a lot of internal modules in text. We keep the upper bound -- inclusive of the latest stable version. - , text >= 1.1.1 && <= 1.2.4.1 + , text >=1.1.1 && <=1.2.5.0 || >=2.0 && <2.1 if flag(dev) ghc-options: -O0 else @@ -110,7 +110,7 @@ QuickCheck >=2.1 && <2.15 , base >=4.7 && <5 , deepseq >=1.1 && <1.5 - , text >=1.1.1 && <1.3 + , text , unicode-transforms if flag(dev) ghc-options: -O0 @@ -129,12 +129,12 @@ test ghc-options: -Wall -fwarn-identities -fwarn-incomplete-record-updates -fwarn-incomplete-uni-patterns -fwarn-tabs build-depends: - QuickCheck >=2.1 && <2.15 - , unicode-data >= 0.1 && <0.2 - , base >=4.7 && <5 - , deepseq >=1.1 && <1.5 - , hspec >= 2.0 && < 3 - , text >=1.1.1 && <1.3 + QuickCheck >= 2.1 && < 2.15 + , unicode-data >= 0.2 && < 0.3 + , base >= 4.7 && < 5 + , deepseq >= 1.1 && < 1.5 + , hspec >= 2.0 && < 3 + , text , unicode-transforms if flag(dev) ghc-options: -O0 @@ -156,9 +156,8 @@ ghc-options: -Wall -fwarn-identities -fwarn-incomplete-record-updates -fwarn-incomplete-uni-patterns -fwarn-tabs build-depends: base >=4.7 && <5 - , getopt-generics >=0.11 && <0.14 , split >=0.1 && <0.3 - , text >=1.1.1 && <1.3 + , text , unicode-transforms if flag(dev) ghc-options: -O0 @@ -180,7 +179,7 @@ , filepath >=1.0 && <2 , path >=0.0.0 && <0.9 , path-io >=0.1.0 && <1.7 - , text >=1.1.1 && <1.3 + , text , unicode-transforms if flag(use-gauge) build-depends: gauge >=0.2.0 && <0.3