master..40a3168 · text

Data/Text.hs+12

since splitsnapshots

unifiedsplit

@@ -60,6 +60,7 @@ module Data.Text

60 , snoc

61 , append

62 , uncons

63+ , unsnoc

63 , head

64 , head

64 , last

65 , last

65 , tail

66 , tail

@@ -545,6 +546,17 @@ init (Text arr off len) | len <= 0 = emptyError "init"

545 unstream (S.init (stream t)) = init t

546 unstream (S.init (stream t)) = init t

546 #-}

547 #-}

547

548

549+-- | /O(1)/ Returns all but the last character and the last character of a

550+-- 'Text', or 'Nothing' if empty.

551+unsnoc :: Text -> Maybe (Text, Char)

552+unsnoc (Text arr off len)

553+ | len <= 0 = Nothing

554+ | n < 0xDC00 || n > 0xDFFF = Just (text arr off (len-1), unsafeChr n)

555+ | otherwise = Just (text arr off (len-2), U16.chr2 n0 n)

556+ where n = A.unsafeIndex arr (off+len-1)

557+ n0 = A.unsafeIndex arr (off+len-2)

558+{-# INLINE [1] unsnoc #-}

559+

548-- | /O(1)/ Tests whether a 'Text' is empty or not. Subject to

560-- | /O(1)/ Tests whether a 'Text' is empty or not. Subject to

549-- fusion.

561-- fusion.

550null :: Text -> Bool

562null :: Text -> Bool

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

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

107import qualified Data.List as L

108import qualified Prelude as P

109import Data.Bits (shiftL)

110−import Data.Char (isLetter)

110+import Data.Char (isLetter, isSpace)

111import Data.Int (Int64)

112import Data.Text.Internal.Fusion.Types

113import 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

462 where

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

464 case next0 s of

465− Done -> Done

465+ Done -> Done

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

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

467 Yield c s'

468− | letter' -> if letter

468+ | nonSpace -> if letter

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

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

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

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

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

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

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

472− where letter' = isLetter c

473+ letter' = isLetter c

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

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

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

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

475

476

476data Justify i s = Just1 !i !s

477data Justify i s = Just1 !i !s

Data/Text/Lazy.hs+10

@@ -68,6 +68,7 @@ module Data.Text.Lazy

68 , snoc

69 , append

70 , uncons

71+ , unsnoc

71 , head

72 , head

72 , last

73 , last

73 , tail

74 , tail

@@ -561,6 +562,15 @@ init Empty = emptyError "init"

561 unstream (S.init (stream t)) = init t

562 unstream (S.init (stream t)) = init t

562 #-}

563 #-}

563

564

565+-- | /O(n\/c)/ Returns the 'init' and 'last' of a 'Text', or 'Nothing' if

566+-- empty.

567+--

568+-- * It is no faster than using 'init' and 'last'.

569+unsnoc :: Text -> Maybe (Text, Char)

570+unsnoc Empty = Nothing

571+unsnoc ts@(Chunk _ _) = Just (init ts, last ts)

572+{-# INLINE unsnoc #-}

573+

564-- | /O(1)/ Tests whether a 'Text' is empty or not. Subject to

574-- | /O(1)/ Tests whether a 'Text' is empty or not. Subject to

565-- fusion.

575-- fusion.

566null :: Text -> Bool

576null :: Text -> Bool

Data/Text/Lazy/Read.hs+3−3

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

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

69

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

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

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

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

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

73−-- is case insensitive.

73+-- This function is case insensitive.

74--

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

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

Data/Text/Read.hs+3−3

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

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

68

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

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

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

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

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

72−-- is case insensitive.

72+-- This function is case insensitive.

73--

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

75-- 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"`.

4+

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.

8+

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

10+ used after some rewrite rule tweaking.

11+

11.2.2.1

121.2.2.1

2

13

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

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

tests/Tests/Properties.hs+23−3

@@ -9,9 +9,9 @@ module Tests.Properties

9 ) where

10

11import Control.Applicative ((<$>), (<*>))

12−import Control.Arrow ((***), second)

12+import Control.Arrow ((***), first, second)

13import 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)

15import Data.Int (Int8, Int16, Int32, Int64)

16import Data.Monoid (Monoid(..))

17import Data.String (IsString(fromString))

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

23import Data.Text.Lazy.Read as TL

24import Data.Text.Read as T

25import Data.Word (Word, Word8, Word16, Word32, Word64)

26+import Data.Maybe (mapMaybe)

26import Numeric (showEFloat, showFFloat, showGFloat, showHex)

27import Numeric (showEFloat, showFFloat, showGFloat, showHex)

27import Prelude hiding (replicate)

28import Prelude hiding (replicate)

28import Test.Framework (Test, testGroup)

29import Test.Framework (Test, testGroup)

@@ -229,6 +230,13 @@ sf_uncons p = (uncons . L.filter p) `eqP`

229 (fmap (second unpackS) . S.uncons . S.filter p)

230 (fmap (second unpackS) . S.uncons . S.filter p)

230t_uncons = uncons `eqP` (fmap (second unpackS) . T.uncons)

231t_uncons = uncons `eqP` (fmap (second unpackS) . T.uncons)

231tl_uncons = uncons `eqP` (fmap (second unpackS) . TL.uncons)

232tl_uncons = uncons `eqP` (fmap (second unpackS) . TL.uncons)

233+

234+unsnoc xs@(_:_) = Just (init xs, last xs)

235+unsnoc [] = Nothing

236+

237+t_unsnoc = unsnoc `eqP` (fmap (first unpackS) . T.unsnoc)

238+tl_unsnoc = unsnoc `eqP` (fmap (first unpackS) . TL.unsnoc)

239+

232s_head = head `eqP` S.head

240s_head = head `eqP` S.head

233sf_head p = (head . L.filter p) `eqP` (S.head . S.filter p)

241sf_head p = (head . L.filter p) `eqP` (S.head . S.filter p)

234t_head = head `eqP` T.head

242t_head = head `eqP` T.head

@@ -326,6 +334,14 @@ t_toUpper_upper t = p (T.toUpper t) >= p t

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

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

327tl_toUpper_upper t = p (TL.toUpper t) >= p t

335tl_toUpper_upper t = p (TL.toUpper t) >= p t

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

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

337+t_toTitle_title t = all (<= 1) (caps t)

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

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

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

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

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

343+ -- in Unicode 7.0

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

329

345

330justifyLeft k c xs = xs ++ L.replicate (k - length xs) c

346justifyLeft k c xs = xs ++ L.replicate (k - length xs) c

331justifyRight m n xs = L.replicate (m - length xs) n ++ xs

347justifyRight m n xs = L.replicate (m - length xs) n ++ xs

@@ -967,6 +983,8 @@ tests =

967 testProperty "sf_uncons" sf_uncons,

983 testProperty "sf_uncons" sf_uncons,

968 testProperty "t_uncons" t_uncons,

984 testProperty "t_uncons" t_uncons,

969 testProperty "tl_uncons" tl_uncons,

985 testProperty "tl_uncons" tl_uncons,

986+ testProperty "t_unsnoc" t_unsnoc,

987+ testProperty "tl_unsnoc" tl_unsnoc,

970 testProperty "s_head" s_head,

988 testProperty "s_head" s_head,

971 testProperty "sf_head" sf_head,

989 testProperty "sf_head" sf_head,

972 testProperty "t_head" t_head,

990 testProperty "t_head" t_head,

@@ -1030,7 +1048,9 @@ tests =

1030 testProperty "tl_toLower_lower" tl_toLower_lower,

1048 testProperty "tl_toLower_lower" tl_toLower_lower,

1031 testProperty "t_toUpper_length" t_toUpper_length,

1049 testProperty "t_toUpper_length" t_toUpper_length,

1032 testProperty "t_toUpper_upper" t_toUpper_upper,

1050 testProperty "t_toUpper_upper" t_toUpper_upper,

1033− testProperty "tl_toUpper_upper" tl_toUpper_upper

1051+ testProperty "tl_toUpper_upper" tl_toUpper_upper,

1052+ testProperty "t_toTitle_title" t_toTitle_title,

1053+ testProperty "t_toTitle_1stNotLower" t_toTitle_1stNotLower

1034 ],

1054 ],

1035

1055

1036 testGroup "justification" [

1056 testGroup "justification" [

tests/text-tests.cabal+17−2

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

20 default: False

21 manual: True

22

23+flag bytestring-builder

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

25+ default: False

26+ manual: False

27+

23executable text-tests

28executable text-tests

24 main-is: Tests.hs

29 main-is: Tests.hs

25

30

@@ -48,7 +53,6 @@ executable text-tests

48 HUnit >= 1.2,

53 HUnit >= 1.2,

49 QuickCheck >= 2.7,

54 QuickCheck >= 2.7,

50 base == 4.*,

55 base == 4.*,

51− bytestring,

52 deepseq,

56 deepseq,

53 directory,

57 directory,

54 quickcheck-unicode,

58 quickcheck-unicode,

@@ -58,6 +62,12 @@ executable text-tests

58 test-framework-quickcheck2 >= 0.2,

62 test-framework-quickcheck2 >= 0.2,

59 text-tests

63 text-tests

60

64

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+

61executable text-tests-stdio

71executable text-tests-stdio

62 main-is: Tests/IO.hs

72 main-is: Tests/IO.hs

63

73

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

137 array,

147 array,

138 base == 4.*,

148 base == 4.*,

139 binary,

149 binary,

140− bytestring,

141 deepseq,

150 deepseq,

142 ghc-prim,

151 ghc-prim,

143 integer-gmp

152 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