master...e825889 · text

git.iko.soy/text

9 files changed+65−24

since splitsnapshots

unifiedsplit

Data/Text.hs+1−1 Data/Text/Foreign.hs+1−1 Data/Text/Internal/Fusion/Common.hs+10−9 Data/Text/Lazy/Read.hs+3−3 Data/Text/Read.hs+3−3 changelog.md+11 tests/Tests/Properties.hs+17−3 tests/text-tests.cabal+18−3 text.cabal+1−1

Data/Text.hs+1−1

@@ -1099,8 +1099,8 @@ takeEnd n t@(Text arr off len)

10991099iterNEnd :: Int -> Text -> Int

11001100iterNEnd n t@(Text _arr _off len) = loop (len-1) n

11011101 where loop i !m

1102+ | m <= 0 = i+1

11021103 | i <= 0 = 0

1103− | m <= 1 = i

11041104 | otherwise = loop (i+d) (m-1)

11051105 where d = reverseIter_ t i

11061106

Data/Text/Foreign.hs+1−1

@@ -108,7 +108,7 @@ takeWord16 (I16 n) t@(Text arr off len)

108108 | n >= len || m >= len = t

109109 | otherwise = Text arr off m

110110 where

111− m | w < 0xDB00 || w > 0xD8FF = n

111+ m | w < 0xD800 || w > 0xDBFF = n

112112 | otherwise = n+1

113113 w = A.unsafeIndex arr (off+n-1)

114114

Data/Text/Internal/Fusion/Common.hs+10−9

@@ -107,7 +107,7 @@ import Prelude (Bool(..), Char, Eq(..), Int, Integral, Maybe(..),

107107import qualified Data.List as L

108108import qualified Prelude as P

109109import Data.Bits (shiftL)

110−import Data.Char (isLetter)

110+import Data.Char (isLetter, isSpace)

111111import Data.Int (Int64)

112112import Data.Text.Internal.Fusion.Types

113113import Data.Text.Internal.Fusion.CaseMapping (foldMapping, lowerMapping, titleMapping,

@@ -462,15 +462,16 @@ toTitle (Stream next0 s0 len) = Stream next (CC (False :*: s0) '\0' '\0') len

462462 where

463463 next (CC (letter :*: s) '\0' _) =

464464 case next0 s of

465− Done -> Done

466− Skip s' -> Skip (CC (letter :*: s') '\0' '\0')

465+ Done -> Done

466+ Skip s' -> Skip (CC (letter :*: s') '\0' '\0')

467467 Yield c s'

468− | letter' -> if letter

469− then lowerMapping c (letter' :*: s')

470− else titleMapping c (letter' :*: s')

471− | otherwise -> Yield c (CC (letter' :*: s') '\0' '\0')

472− where letter' = isLetter c

473− next (CC s a b) = Yield a (CC s b '\0')

468+ | nonSpace -> if letter

469+ then lowerMapping c (nonSpace :*: s')

470+ else titleMapping c (letter' :*: s')

471+ | otherwise -> Yield c (CC (letter' :*: s') '\0' '\0')

472+ where nonSpace = P.not (isSpace c)

473+ letter' = isLetter c

474+ next (CC s a b) = Yield a (CC s b '\0')

474475{-# INLINE [0] toTitle #-}

475476

476477data Justify i s = Just1 !i !s

Data/Text/Lazy/Read.hs+3−3

@@ -68,9 +68,9 @@ decimal txt

6868 go n d = (n * 10 + fromIntegral (digitToInt d))

6969

7070-- | Read a hexadecimal integer, consisting of an optional leading

71−-- @\"0x\"@ followed by at least one decimal digit. Input is consumed

72−-- until a non-hex-digit or end of string is reached. This function

73−-- is case insensitive.

71+-- @\"0x\"@ followed by at least one hexadecimal digit. Input is

72+-- consumed until a non-hex-digit or end of string is reached.

73+-- This function is case insensitive.

7474--

7575-- This function does not handle leading sign characters. If you need

7676-- to handle signed input, use @'signed' 'hexadecimal'@.

Data/Text/Read.hs+3−3

@@ -67,9 +67,9 @@ decimal txt

6767 go n d = (n * 10 + fromIntegral (digitToInt d))

6868

6969-- | Read a hexadecimal integer, consisting of an optional leading

70−-- @\"0x\"@ followed by at least one decimal digit. Input is consumed

71−-- until a non-hex-digit or end of string is reached. This function

72−-- is case insensitive.

70+-- @\"0x\"@ followed by at least one hexadecimal digit. Input is

71+-- consumed until a non-hex-digit or end of string is reached.

72+-- This function is case insensitive.

7373--

7474-- This function does not handle leading sign characters. If you need

7575-- to handle signed input, use @'signed' 'hexadecimal'@.

changelog.md+11

@@ -1,3 +1,14 @@

1+* The `toTitle` function now correctly handles letters that

2+ immediately follow puncutation. Before, `"there's"` would turn into

3+ `"There'S"`. Now, it becomes `"There's"`.

5+* The implementation of unstreaming is faster, resulting in operations

6+ such as `map` and `intersperse` speeding up by up to 30%, with

7+ smaller code generated.

9+* The optimised length comparison function is now more likely to be

10+ used after some rewrite rule tweaking.

11+

1121.2.2.1

213

314* The switch to `integer-pure` in 1.2.2.0 was apparently mistaken.

tests/Tests/Properties.hs+17−3

@@ -11,7 +11,7 @@ module Tests.Properties

1111import Control.Applicative ((<$>), (<*>))

1212import Control.Arrow ((***), second)

1313import Data.Bits ((.&.))

14−import Data.Char (chr, isDigit, isHexDigit, isLower, isSpace, isUpper, ord)

14+import Data.Char (chr, isDigit, isHexDigit, isLower, isSpace, isLetter, isUpper, ord)

1515import Data.Int (Int8, Int16, Int32, Int64)

1616import Data.Monoid (Monoid(..))

1717import Data.String (IsString(fromString))

@@ -23,6 +23,7 @@ import Data.Text.Internal.Search (indices)

2323import Data.Text.Lazy.Read as TL

2424import Data.Text.Read as T

2525import Data.Word (Word, Word8, Word16, Word32, Word64)

26+import Data.Maybe (mapMaybe)

2627import Numeric (showEFloat, showFFloat, showGFloat, showHex)

2728import Prelude hiding (replicate)

2829import Test.Framework (Test, testGroup)

@@ -37,6 +38,7 @@ import qualified Control.Exception as Exception

3738import qualified Data.Bits as Bits (shiftL, shiftR)

3839import qualified Data.ByteString as B

3940import qualified Data.ByteString.Lazy as BL

41+import qualified Data.Char as C

4042import qualified Data.List as L

4143import qualified Data.Text as T

4244import qualified Data.Text.Encoding as E

@@ -326,6 +328,16 @@ t_toUpper_upper t = p (T.toUpper t) >= p t

326328 where p = T.length . T.filter isUpper

327329tl_toUpper_upper t = p (TL.toUpper t) >= p t

328330 where p = TL.length . TL.filter isUpper

331+t_toTitle_title t = all (<= 1) (caps w)

332+ where caps = fmap (T.length . T.filter isUpper) . T.words . T.toTitle

333+ -- TIL: there exist uppercase-only letters

334+ w = T.filter (\c -> if C.isUpper c then C.toLower c /= c else True) t

335+t_toTitle_1stNotLower = and . notLow . T.toTitle . T.filter stable

336+ where notLow = mapMaybe (fmap (not . isLower) . (T.find isLetter)) . T.words

337+ -- Surprise! The Spanish/Portuguese ordinal indicators changed

338+ -- from category Ll (letter, lowercase) to Lo (letter, other)

339+ -- in Unicode 7.0

340+ stable c = c /= '\170' && c /= '\186'

329341

330342justifyLeft k c xs = xs ++ L.replicate (k - length xs) c

331343justifyRight m n xs = L.replicate (m - length xs) n ++ xs

@@ -867,7 +879,7 @@ tl_write_read_line e m b t = write_read head TL.filter TL.hPutStrLn

867879

868880t_dropWord16 m t = dropWord16 m t `T.isSuffixOf` t

869881t_takeWord16 m t = takeWord16 m t `T.isPrefixOf` t

870−t_take_drop_16 m t = T.append (takeWord16 n t) (dropWord16 n t) == t

882+t_take_drop_16 m t = T.append (takeWord16 n t) (dropWord16 n t) === t

871883 where n = small m

872884t_use_from t = monadicIO $ assert . (==t) =<< run (useAsPtr t fromPtr)

873885

@@ -1030,7 +1042,9 @@ tests =

10301042 testProperty "tl_toLower_lower" tl_toLower_lower,

10311043 testProperty "t_toUpper_length" t_toUpper_length,

10321044 testProperty "t_toUpper_upper" t_toUpper_upper,

1033− testProperty "tl_toUpper_upper" tl_toUpper_upper

1045+ testProperty "tl_toUpper_upper" tl_toUpper_upper,

1046+ testProperty "t_toTitle_title" t_toTitle_title,

1047+ testProperty "t_toTitle_1stNotLower" t_toTitle_1stNotLower

10341048 ],

10351049

10361050 testGroup "justification" [

tests/text-tests.cabal+18−3

@@ -20,6 +20,11 @@ flag hpc

2020 default: False

2121 manual: True

2222

23+flag bytestring-builder

24+ description: Depend on the bytestring-builder package for backwards compatibility.

25+ default: False

26+ manual: False

27+

2328executable text-tests

2429 main-is: Tests.hs

2530

@@ -48,16 +53,21 @@ executable text-tests

4853 HUnit >= 1.2,

4954 QuickCheck >= 2.7,

5055 base == 4.*,

51− bytestring,

5256 deepseq,

5357 directory,

54− quickcheck-unicode,

58+ quickcheck-unicode >= 1.0.1.0,

5559 random,

5660 test-framework >= 0.4,

5761 test-framework-hunit >= 0.2,

5862 test-framework-quickcheck2 >= 0.2,

5963 text-tests

6064

65+ if flag(bytestring-builder)

66+ build-depends: bytestring >= 0.9 && < 0.10.4,

67+ bytestring-builder >= 0.10.4

68+ else

69+ build-depends: bytestring >= 0.10.4

70+

6171executable text-tests-stdio

6272 main-is: Tests/IO.hs

6373

@@ -137,7 +147,12 @@ library

137147 array,

138148 base == 4.*,

139149 binary,

140− bytestring,

141150 deepseq,

142151 ghc-prim,

143152 integer-gmp

153+

154+ if flag(bytestring-builder)

155+ build-depends: bytestring >= 0.9 && < 0.10.4,

156+ bytestring-builder >= 0.10.4

157+ else

158+ build-depends: bytestring >= 0.10.4

text.cabal+1−1

@@ -180,7 +180,7 @@ test-suite tests

180180 deepseq,

181181 directory,

182182 ghc-prim,

183− quickcheck-unicode,

183+ quickcheck-unicode >= 1.0.1.0,

184184 random,

185185 test-framework >= 0.4,

186186 test-framework-hunit >= 0.2,