@@ -3,6 +3,9 @@
33#if __GLASGOW_HASKELL__ >= 702
44{-# LANGUAGE Trustworthy #-}
55#endif
6+#if __GLASGOW_HASKELL__ >= 708
7+{-# LANGUAGE TypeFamilies #-}
8+#endif
69
710-- |
811-- Module : Data.Text
@@ -29,8 +32,8 @@
2932--
3033-- To use an extended and very rich family of functions for working
3134-- with Unicode text (including normalization, regular expressions,
32−-- non-standard encodings, text breaking, and locales), see the
33−-- @text-icu@ package: <http://hackage.haskell.org/package/text-icu>
35+-- non-standard encodings, text breaking, and locales), see
36+-- <http://hackage.haskell.org/package/text-icu the text-icu package >.
3437
3538module Data.Text
3639 (
@@ -122,8 +125,11 @@ module Data.Text
122125
123126 -- ** Breaking strings
124127 , take
128+ , takeEnd
125129 , drop
130+ , dropEnd
126131 , takeWhile
132+ , takeWhileEnd
127133 , dropWhile
128134 , dropWhileEnd
129135 , dropAround
@@ -186,35 +192,43 @@ module Data.Text
186192
187193 -- * Low level operations
188194 , copy
195+ , unpackCString#
189196 ) where
190197
191198import Prelude (Char, Bool(..), Int, Maybe(..), String,
192199 Eq(..), Ord(..), Ordering(..), (++),
193− Read(..), Show(..),
194− (&&), (||), (+), (-), (.), ($), ($!), (>>), (*),
195− maxBound, not, return, otherwise, quot)
200+ Read(..),
201+ (&&), (||), (+), (-), (.), ($), ($!), (>>),
202+ not, return, otherwise, quot)
196203#if defined(HAVE_DEEPSEQ)
197−import Control.DeepSeq (NFData)
204+import Control.DeepSeq (NFData(rnf))
198205#endif
199206#if defined(ASSERTS)
200207import Control.Exception (assert)
201208#endif
202209import Data.Char (isSpace)
203−import Data.Data (Data(gfoldl, toConstr, gunfold, dataTypeOf))
204−import Data.Data (mkNoRepType)
210+import Data.Data (Data(gfoldl, toConstr, gunfold, dataTypeOf), constrIndex,
211+ Constr, mkConstr, DataType, mkDataType, Fixity(Prefix))
205212import Control.Monad (foldM)
213+import Control.Monad.ST (ST)
206214import qualified Data.Text.Array as A
207215import qualified Data.List as L
216+import Data.Binary (Binary(get, put))
208217import Data.Monoid (Monoid(..))
218+#if MIN_VERSION_base(4,9,0)
219+import Data.Semigroup (Semigroup(..))
220+#endif
209221import Data.String (IsString(..))
210222import qualified Data.Text.Internal.Fusion as S
211223import qualified Data.Text.Internal.Fusion.Common as S
224+import Data.Text.Encoding (decodeUtf8', encodeUtf8)
212225import Data.Text.Internal.Fusion (stream, reverseStream, unstream)
213226import Data.Text.Internal.Private (span_)
214−import Data.Text.Internal (Text(..), empty, firstf, safe, text, textP)
227+import Data.Text.Internal (Text(..), empty, firstf, mul, safe, text)
228+import Data.Text.Show (singleton, unpack, unpackCString#)
215229import qualified Prelude as P
216230import Data.Text.Unsafe (Iter(..), iter, iter_, lengthWord16, reverseIter,
217− unsafeHead, unsafeTail)
231+ reverseIter_, unsafeHead, unsafeTail)
218232import Data.Text.Internal.Unsafe.Char (unsafeChr)
219233import qualified Data.Text.Internal.Functions as F
220234import qualified Data.Text.Internal.Encoding.Utf16 as U16
@@ -224,12 +238,13 @@ import Data.ByteString (ByteString)
224238import qualified Data.Text.Lazy as L
225239import Data.Int (Int64)
226240#endif
227−#if __GLASGOW_HASKELL__ >= 702
228−import qualified GHC.CString as GHC
229−#else
230−import qualified GHC.Base as GHC
241+import GHC.Base (eqInt, neInt, gtInt, geInt, ltInt, leInt)
242+#if __GLASGOW_HASKELL__ >= 708
243+import qualified GHC.Exts as Exts
244+#endif
245+#if MIN_VERSION_base(4,7,0)
246+import Text.Printf (PrintfArg, formatArg, formatString)
231247#endif
232−import GHC.Prim (Addr#)
233248
234249-- $strict
235250--
@@ -239,29 +254,32 @@ import GHC.Prim (Addr#)
239254-- lazy @Text@ type consists of a list of strict chunks.
240255--
241256-- The strict 'Text' type requires that an entire string fit into
242−-- memory at once. The lazy @Text@ type is capable of streaming
243−-- strings that are larger than memory using a small memory footprint.
244−-- In many cases, the overhead of chunked streaming makes the lazy
245−-- @Text@ type slower than its strict counterpart, but this is not
246−-- always the case. Sometimes, the time complexity of a function in
247−-- one module may be different from the other, due to their differing
248−-- internal structures.
257+-- memory at once. The lazy 'Data.Text.Lazy.Text' type is capable of
258+-- streaming strings that are larger than memory using a small memory
259+-- footprint. In many cases, the overhead of chunked streaming makes
260+-- the lazy 'Data.Text.Lazy.Text' type slower than its strict
261+-- counterpart, but this is not always the case. Sometimes, the time
262+-- complexity of a function in one module may be different from the
263+-- other, due to their differing internal structures.
249264--
250265-- Each module provides an almost identical API, with the main
251266-- difference being that the strict module uses 'Int' values for
252−-- lengths and counts, while the lazy module uses 'Int64' lengths.
267+-- lengths and counts, while the lazy module uses 'Data.Int.Int64'
268+-- lengths.
253269
254270-- $replacement
255271--
256272-- A 'Text' value is a sequence of Unicode scalar values, as defined
257−-- in §3.9, definition D76 of the Unicode 5.2 standard:
258−-- <http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=35>. As
259−-- such, a 'Text' cannot contain values in the range U+D800 to U+DFFF
260−-- inclusive. Haskell implementations admit all Unicode code points
261−-- (§3.4, definition D10) as 'Char' values, including code points
262−-- from this invalid range. This means that there are some 'Char'
263−-- values that are not valid Unicode scalar values, and the functions
264−-- in this module must handle those cases.
273+-- in
274+-- <http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=35 §3.9, definition D76 of the Unicode 5.2 standard >.
275+-- As such, a 'Text' cannot contain values in the range U+D800 to
276+-- U+DFFF inclusive. Haskell implementations admit all Unicode code
277+-- points
278+-- (<http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=13 §3.4, definition D10 >)
279+-- as 'Char' values, including code points from this invalid range.
280+-- This means that there are some 'Char' values that are not valid
281+-- Unicode scalar values, and the functions in this module must handle
282+-- those cases.
265283--
266284-- Within this module, many functions construct a 'Text' from one or
267285-- more 'Char' values. Those functions will substitute 'Char' values
@@ -275,8 +293,8 @@ import GHC.Prim (Addr#)
275293-- range U+D800 through U+DFFF are used by UTF-16 to denote surrogate
276294-- code points, and so cannot be represented. The functions replace
277295-- invalid scalar values, instead of dropping them, as a security
278−-- measure. For details, see Unicode Technical Report 36, §3.5:
279−-- <http://unicode.org/reports/tr36/#Deletion_of_Noncharacters>)
296+-- measure. For details, see
297+-- <http://unicode.org/reports/tr36/#Deletion_of_Noncharacters Unicode Technical Report 36, §3.5 >.)
280298
281299-- $fusion
282300--
@@ -294,11 +312,11 @@ import GHC.Prim (Addr#)
294312-- > countChars = T.length . T.toUpper . E.decodeUtf8
295313--
296314-- From the type signatures involved, this looks like it should
297−-- allocate one 'ByteString' value, and two 'Text' values. However,
298−-- when a module is compiled with optimisation enabled under GHC, the
299−-- two intermediate 'Text' values will be optimised away, and the
300−-- function will be compiled down to a single loop over the source
301−-- 'ByteString'.
315+-- allocate one 'Data.ByteString.ByteString' value, and two 'Text'
316+-- values. However, when a module is compiled with optimisation
317+-- enabled under GHC, the two intermediate 'Text' values will be
318+-- optimised away, and the function will be compiled down to a single
319+-- loop over the source 'Data.ByteString.ByteString'.
302320--
303321-- Functions that can be fused by the compiler are documented with the
304322-- phrase \"Subject to fusion\".
@@ -312,41 +330,82 @@ instance Eq Text where
312330instance Ord Text where
313331 compare = compareText
314332
315−instance Show Text where
316− showsPrec p ps r = showsPrec p (unpack ps) r
317−
318333instance Read Text where
319334 readsPrec p str = [(pack x,y) | (x,y) <- readsPrec p str]
320335
336+#if MIN_VERSION_base(4,9,0)
337+-- Semigroup orphan instances for older GHCs are provided by
338+-- 'semigroups` package
339+
340+instance Semigroup Text where
341+ (<>) = append
342+#endif
343+
321344instance Monoid Text where
322345 mempty = empty
346+#if MIN_VERSION_base(4,9,0)
347+ mappend = (<>) -- future-proof definition
348+#else
323349 mappend = append
350+#endif
324351 mconcat = concat
325352
326353instance IsString Text where
327354 fromString = pack
328355
356+#if __GLASGOW_HASKELL__ >= 708
357+instance Exts.IsList Text where
358+ type Item Text = Char
359+ fromList = pack
360+ toList = unpack
361+#endif
362+
329363#if defined(HAVE_DEEPSEQ)
330−instance NFData Text
364+instance NFData Text where rnf !_ = ()
331365#endif
332366
367+instance Binary Text where
368+ put t = put (encodeUtf8 t)
369+ get = do
370+ bs <- get
371+ case decodeUtf8' bs of
372+ P.Left exn -> P.fail (P.show exn)
373+ P.Right a -> P.return a
374+
333375-- | This instance preserves data abstraction at the cost of inefficiency.
334376-- We omit reflection services for the sake of data abstraction.
335377--
336−-- This instance was created by copying the behavior of Data.Set and
337−-- Data.Map. If you feel a mistake has been made, please feel free to
338−-- submit improvements.
378+-- This instance was created by copying the updated behavior of
379+-- @"Data.Set".@'Data.Set.Set' and @"Data.Map".@'Data.Map.Map'. If you
380+-- feel a mistake has been made, please feel free to submit
381+-- improvements.
339382--
340−-- Original discussion is archived here:
383+-- The original discussion is archived here:
384+-- <http://groups.google.com/group/haskell-cafe/browse_thread/thread/b5bbb1b28a7e525d/0639d46852575b93 could we get a Data instance for Data.Text.Text? >
341385--
342−-- "could we get a Data instance for Data.Text.Text?"
343−-- http://groups.google.com/group/haskell-cafe/browse_thread/thread/b5bbb1b28a7e525d/0639d46852575b93
386+-- The followup discussion that changed the behavior of 'Data.Set.Set'
387+-- and 'Data.Map.Map' is archived here:
388+-- <http://markmail.org/message/trovdc6zkphyi3cr#query:+page:1+mid:a46der3iacwjcf6n+state:results Proposal: Allow gunfold for Data.Map, ... >
344389
345390instance Data Text where
346391 gfoldl f z txt = z pack `f` (unpack txt)
347− toConstr _ = P.error "Data.Text.Text.toConstr"
348− gunfold _ _ = P.error "Data.Text.Text.gunfold"
349− dataTypeOf _ = mkNoRepType "Data.Text.Text"
392+ toConstr _ = packConstr
393+ gunfold k z c = case constrIndex c of
394+ 1 -> k (z pack)
395+ _ -> P.error "gunfold"
396+ dataTypeOf _ = textDataType
397+
398+#if MIN_VERSION_base(4,7,0)
399+-- | Only defined for @base-4.7.0.0@ and later
400+instance PrintfArg Text where
401+ formatArg txt = formatString $ unpack txt
402+#endif
403+
404+packConstr :: Constr
405+packConstr = mkConstr textDataType "pack" [] Prefix
406+
407+textDataType :: DataType
408+textDataType = mkDataType "Data.Text.Text" [packConstr]
350409
351410-- | /O(n)/ Compare two 'Text' values lexicographically.
352411compareText :: Text -> Text -> Ordering
@@ -371,30 +430,6 @@ pack :: String -> Text
371430pack = unstream . S.map safe . S.streamList
372431{-# INLINE [1] pack #-}
373432
374−-- | /O(n)/ Convert a 'Text' into a 'String'. Subject to fusion.
375−unpack :: Text -> String
376−unpack = S.unstreamList . stream
377−{-# INLINE [1] unpack #-}
378−
379−-- | /O(n)/ Convert a literal string into a Text. Subject to fusion.
380−unpackCString# :: Addr# -> Text
381−unpackCString# addr# = unstream (S.streamCString# addr#)
382−{-# NOINLINE unpackCString# #-}
383−
384−{-# RULES "TEXT literal" forall a.
385− unstream (S.map safe (S.streamList (GHC.unpackCString# a)))
386− = unpackCString# a #-}
387−
388−{-# RULES "TEXT literal UTF8" forall a.
389− unstream (S.map safe (S.streamList (GHC.unpackCStringUtf8# a)))
390− = unpackCString# a #-}
391−
392−-- | /O(1)/ Convert a character into a Text. Subject to fusion.
393−-- Performs replacement on invalid scalar values.
394−singleton :: Char -> Text
395−singleton = unstream . S.singleton . safe
396−{-# INLINE [1] singleton #-}
397−
398433-- -----------------------------------------------------------------------------
399434-- * Basic functions
400435
@@ -425,6 +460,7 @@ append a@(Text arr1 off1 len1) b@(Text arr2 off2 len2)
425460 | otherwise = overflowError "append"
426461 where
427462 len = len1+len2
463+ x :: ST s (A.MArray s)
428464 x = do
429465 arr <- A.new len
430466 A.copyI arr 0 arr1 off1 len1
@@ -450,8 +486,8 @@ head t = S.head (stream t)
450486uncons :: Text -> Maybe (Char, Text)
451487uncons t@(Text arr off len)
452488 | len <= 0 = Nothing
453− | otherwise = Just (c, textP arr (off+d) (len-d))
454− where Iter c d = iter t 0
489+ | otherwise = Just $ let !(Iter c d) = iter t 0
490+ in (c, text arr (off+d) (len-d))
455491{-# INLINE [1] uncons #-}
456492
457493-- | Lifted from Control.Arrow and specialized.
@@ -481,7 +517,7 @@ last (Text arr off len)
481517tail :: Text -> Text
482518tail t@(Text arr off len)
483519 | len <= 0 = emptyError "tail"
484− | otherwise = textP arr (off+d) (len-d)
520+ | otherwise = text arr (off+d) (len-d)
485521 where d = iter_ t 0
486522{-# INLINE [1] tail #-}
487523
@@ -496,8 +532,8 @@ tail t@(Text arr off len)
496532-- be non-empty. Subject to fusion.
497533init :: Text -> Text
498534init (Text arr off len) | len <= 0 = emptyError "init"
499− | n >= 0xDC00 && n <= 0xDFFF = textP arr off (len-2)
500− | otherwise = textP arr off (len-1)
535+ | n >= 0xDC00 && n <= 0xDFFF = text arr off (len-2)
536+ | otherwise = text arr off (len-1)
501537 where
502538 n = A.unsafeIndex arr (off+len-1)
503539{-# INLINE [1] init #-}
@@ -536,7 +572,9 @@ isSingleton = S.isSingleton . stream
536572-- Subject to fusion.
537573length :: Text -> Int
538574length t = S.length (stream t)
539−{-# INLINE length #-}
575+{-# INLINE [0] length #-}
576+-- length needs to be phased after the compareN/length rules otherwise
577+-- it may inline before the rules have an opportunity to fire.
540578
541579-- | /O(n)/ Compare the count of characters in a 'Text' to a number.
542580-- Subject to fusion.
@@ -555,32 +593,32 @@ compareLength t n = S.compareLengthI (stream t) n
555593
556594{-# RULES
557595"TEXT ==N/length -> compareLength/==EQ" [~1] forall t n.
558− (==) (length t) n = compareLength t n == EQ
596+ eqInt (length t) n = compareLength t n == EQ
559597 #-}
560598
561599{-# RULES
562600"TEXT /=N/length -> compareLength//=EQ" [~1] forall t n.
563− (/=) (length t) n = compareLength t n /= EQ
601+ neInt (length t) n = compareLength t n /= EQ
564602 #-}
565603
566604{-# RULES
567605"TEXT <N/length -> compareLength/==LT" [~1] forall t n.
568− (<) (length t) n = compareLength t n == LT
606+ ltInt (length t) n = compareLength t n == LT
569607 #-}
570608
571609{-# RULES
572610"TEXT <=N/length -> compareLength//=GT" [~1] forall t n.
573− (<=) (length t) n = compareLength t n /= GT
611+ leInt (length t) n = compareLength t n /= GT
574612 #-}
575613
576614{-# RULES
577615"TEXT >N/length -> compareLength/==GT" [~1] forall t n.
578− (>) (length t) n = compareLength t n == GT
616+ gtInt (length t) n = compareLength t n == GT
579617 #-}
580618
581619{-# RULES
582620"TEXT >=N/length -> compareLength//=LT" [~1] forall t n.
583− (>=) (length t) n = compareLength t n /= LT
621+ geInt (length t) n = compareLength t n /= LT
584622 #-}
585623
586624-- -----------------------------------------------------------------------------
@@ -611,22 +649,48 @@ reverse :: Text -> Text
611649reverse t = S.reverse (stream t)
612650{-# INLINE reverse #-}
613651
614−-- | /O(m+n)/ Replace every occurrence of one substring with another.
652+-- | /O(m+n)/ Replace every non-overlapping occurrence of @needle@ in
653+-- @haystack@ with @replacement@.
654+--
655+-- This function behaves as though it was defined as follows:
656+--
657+-- @
658+-- replace needle replacement haystack =
659+-- 'intercalate' replacement ('splitOn' needle haystack)
660+-- @
661+--
662+-- As this suggests, each occurrence is replaced exactly once. So if
663+-- @needle@ occurs in @replacement@, that occurrence will /not/ itself
664+-- be replaced recursively:
665+--
666+-- > replace "oo" "foo" "oo" == "foo"
667+--
668+-- In cases where several instances of @needle@ overlap, only the
669+-- first one will be replaced:
670+--
671+-- > replace "ofo" "bar" "ofofo" == "barfo"
615672--
616673-- In (unlikely) bad cases, this function's time complexity degrades
617674-- towards /O(n*m)/.
618−replace :: Text -> Text -> Text -> Text
675+replace :: Text
676+ -- ^ @needle@ to search for. If this string is empty, an
677+ -- error will occur.
678+ -> Text
679+ -- ^ @replacement@ to replace @needle@ with.
680+ -> Text
681+ -- ^ @haystack@ in which to search.
682+ -> Text
619683replace needle@(Text _ _ neeLen)
620684 (Text repArr repOff repLen)
621685 haystack@(Text hayArr hayOff hayLen)
622686 | neeLen == 0 = emptyError "replace"
623687 | L.null ixs = haystack
624688 | len > 0 = Text (A.run x) 0 len
625− | len < 0 = overflowError "replace"
626689 | otherwise = empty
627690 where
628691 ixs = indices needle haystack
629− len = hayLen - (neeLen - repLen) * L.length ixs
692+ len = hayLen - (neeLen - repLen) `mul` L.length ixs
693+ x :: ST s (A.MArray s)
630694 x = do
631695 marr <- A.new len
632696 let loop (i:is) o d = do
@@ -656,9 +720,8 @@ replace needle@(Text _ _ neeLen)
656720-- /Note/: In some languages, case conversion is a locale- and
657721-- context-dependent operation. The case conversion functions in this
658722-- module are /not/ locale sensitive. Programs that require locale
659−-- sensitivity should use appropriate versions of the case mapping
660−-- functions from the @text-icu@ package:
661−-- <http://hackage.haskell.org/package/text-icu>
723+-- sensitivity should use appropriate versions of the
724+-- <http://hackage.haskell.org/package/text-icu-0.6.3.7/docs/Data-Text-ICU.html#g:4 case mapping functions from the text-icu package >.
662725
663726-- | /O(n)/ Convert a string to folded case. Subject to fusion.
664727--
@@ -678,7 +741,7 @@ replace needle@(Text _ _ neeLen)
678741-- instead of itself.
679742toCaseFold :: Text -> Text
680743toCaseFold t = unstream (S.toCaseFold (stream t))
681−{-# INLINE [0] toCaseFold #-}
744+{-# INLINE toCaseFold #-}
682745
683746-- | /O(n)/ Convert a string to lower case, using simple case
684747-- conversion. Subject to fusion.
@@ -838,6 +901,7 @@ concat ts = case ts' of
838901 where
839902 ts' = L.filter (not . null) ts
840903 len = sumP "concat" $ L.map lengthWord16 ts'
904+ go :: ST s (A.MArray s)
841905 go = do
842906 arr <- A.new len
843907 let step i (Text a o l) =
@@ -851,13 +915,13 @@ concatMap f = concat . foldr ((:) . f) []
851915{-# INLINE concatMap #-}
852916
853917-- | /O(n)/ 'any' @p@ @t@ determines whether any character in the
854−-- 'Text' @t@ satisifes the predicate @p@. Subject to fusion.
918+-- 'Text' @t@ satisfies the predicate @p@. Subject to fusion.
855919any :: (Char -> Bool) -> Text -> Bool
856920any p t = S.any p (stream t)
857921{-# INLINE any #-}
858922
859923-- | /O(n)/ 'all' @p@ @t@ determines whether all characters in the
860−-- 'Text' @t@ satisify the predicate @p@. Subject to fusion.
924+-- 'Text' @t@ satisfy the predicate @p@. Subject to fusion.
861925all :: (Char -> Bool) -> Text -> Bool
862926all p t = S.all p (stream t)
863927{-# INLINE all #-}
@@ -948,10 +1012,10 @@ replicate n t@(Text a o l)
9481012 | n <= 0 || l <= 0 = empty
9491013 | n == 1 = t
9501014 | isSingleton t = replicateChar n (unsafeHead t)
951− | n <= maxBound `quot` l = Text (A.run x) 0 len
952− | otherwise = overflowError "replicate"
1015+ | otherwise = Text (A.run x) 0 len
9531016 where
954− len = l * n
1017+ len = l `mul` n
1018+ x :: ST s (A.MArray s)
9551019 x = do
9561020 arr <- A.new len
9571021 let loop !d !i | i >= n = return arr
@@ -1002,14 +1066,16 @@ take :: Int -> Text -> Text
10021066take n t@(Text arr off len)
10031067 | n <= 0 = empty
10041068 | n >= len = t
1005− | otherwise = Text arr off (loop 0 0)
1006− where
1007− loop !i !cnt
1008− | i >= len || cnt >= n = i
1009− | otherwise = loop (i+d) (cnt+1)
1010− where d = iter_ t i
1069+ | otherwise = text arr off (iterN n t)
10111070{-# INLINE [1] take #-}
10121071
1072+iterN :: Int -> Text -> Int
1073+iterN n t@(Text _arr _off len) = loop 0 0
1074+ where loop !i !cnt
1075+ | i >= len || cnt >= n = i
1076+ | otherwise = loop (i+d) (cnt+1)
1077+ where d = iter_ t i
1078+
10131079{-# RULES
10141080"TEXT take -> fused" [~1] forall n t.
10151081 take n t = unstream (S.take n (stream t))
@@ -1017,6 +1083,27 @@ take n t@(Text arr off len)
10171083 unstream (S.take n (stream t)) = take n t
10181084 #-}
10191085
1086+-- | /O(n)/ 'takeEnd' @n@ @t@ returns the suffix remaining after
1087+-- taking @n@ characters from the end of @t@.
1088+--
1089+-- Examples:
1090+--
1091+-- > takeEnd 3 "foobar" == "bar"
1092+takeEnd :: Int -> Text -> Text
1093+takeEnd n t@(Text arr off len)
1094+ | n <= 0 = empty
1095+ | n >= len = t
1096+ | otherwise = text arr (off+i) (len-i)
1097+ where i = iterNEnd n t
1098+
1099+iterNEnd :: Int -> Text -> Int
1100+iterNEnd n t@(Text _arr _off len) = loop (len-1) n
1101+ where loop i !m
1102+ | i <= 0 = 0
1103+ | m <= 1 = i
1104+ | otherwise = loop (i+d) (m-1)
1105+ where d = reverseIter_ t i
1106+
10201107-- | /O(n)/ 'drop' @n@, applied to a 'Text', returns the suffix of the
10211108-- 'Text' after the first @n@ characters, or the empty 'Text' if @n@
10221109-- is greater than the length of the 'Text'. Subject to fusion.
@@ -1024,11 +1111,8 @@ drop :: Int -> Text -> Text
10241111drop n t@(Text arr off len)
10251112 | n <= 0 = t
10261113 | n >= len = empty
1027− | otherwise = loop 0 0
1028− where loop !i !cnt
1029− | i >= len || cnt >= n = Text arr (off+i) (len-i)
1030− | otherwise = loop (i+d) (cnt+1)
1031− where d = iter_ t i
1114+ | otherwise = text arr (off+i) (len-i)
1115+ where i = iterN n t
10321116{-# INLINE [1] drop #-}
10331117
10341118{-# RULES
@@ -1038,6 +1122,18 @@ drop n t@(Text arr off len)
10381122 unstream (S.drop n (stream t)) = drop n t
10391123 #-}
10401124
1125+-- | /O(n)/ 'dropEnd' @n@ @t@ returns the prefix remaining after
1126+-- dropping @n@ characters from the end of @t@.
1127+--
1128+-- Examples:
1129+--
1130+-- > dropEnd 3 "foobar" == "foo"
1131+dropEnd :: Int -> Text -> Text
1132+dropEnd n t@(Text arr off len)
1133+ | n <= 0 = t
1134+ | n >= len = empty
1135+ | otherwise = text arr off (iterNEnd n t)
1136+
10411137-- | /O(n)/ 'takeWhile', applied to a predicate @p@ and a 'Text',
10421138-- returns the longest prefix (possibly empty) of elements that
10431139-- satisfy @p@. Subject to fusion.
@@ -1045,7 +1141,7 @@ takeWhile :: (Char -> Bool) -> Text -> Text
10451141takeWhile p t@(Text arr off len) = loop 0
10461142 where loop !i | i >= len = t
10471143 | p c = loop (i+d)
1048− | otherwise = textP arr off i
1144+ | otherwise = text arr off i
10491145 where Iter c d = iter t i
10501146{-# INLINE [1] takeWhile #-}
10511147
@@ -1056,6 +1152,27 @@ takeWhile p t@(Text arr off len) = loop 0
10561152 unstream (S.takeWhile p (stream t)) = takeWhile p t
10571153 #-}
10581154
1155+-- | /O(n)/ 'takeWhileEnd', applied to a predicate @p@ and a 'Text',
1156+-- returns the longest suffix (possibly empty) of elements that
1157+-- satisfy @p@. Subject to fusion.
1158+-- Examples:
1159+--
1160+-- > takeWhileEnd (=='o') "foo" == "oo"
1161+takeWhileEnd :: (Char -> Bool) -> Text -> Text
1162+takeWhileEnd p t@(Text arr off len) = loop (len-1) len
1163+ where loop !i !l | l <= 0 = t
1164+ | p c = loop (i+d) (l+d)
1165+ | otherwise = text arr (off+l) (len-l)
1166+ where (c,d) = reverseIter t i
1167+{-# INLINE [1] takeWhileEnd #-}
1168+
1169+{-# RULES
1170+"TEXT takeWhileEnd -> fused" [~1] forall p t.
1171+ takeWhileEnd p t = S.reverse (S.takeWhile p (S.reverseStream t))
1172+"TEXT takeWhileEnd -> unfused" [1] forall p t.
1173+ S.reverse (S.takeWhile p (S.reverseStream t)) = takeWhileEnd p t
1174+ #-}
1175+
10591176-- | /O(n)/ 'dropWhile' @p@ @t@ returns the suffix remaining after
10601177-- 'takeWhile' @p@ @t@. Subject to fusion.
10611178dropWhile :: (Char -> Bool) -> Text -> Text
@@ -1074,8 +1191,9 @@ dropWhile p t@(Text arr off len) = loop 0 0
10741191 #-}
10751192
10761193-- | /O(n)/ 'dropWhileEnd' @p@ @t@ returns the prefix remaining after
1077−-- dropping characters that fail the predicate @p@ from the end of
1194+-- dropping characters that satisfy the predicate @p@ from the end of
10781195-- @t@. Subject to fusion.
1196+--
10791197-- Examples:
10801198--
10811199-- > dropWhileEnd (=='.') "foo..." == "foo"
@@ -1095,7 +1213,7 @@ dropWhileEnd p t@(Text arr off len) = loop (len-1) len
10951213 #-}
10961214
10971215-- | /O(n)/ 'dropAround' @p@ @t@ returns the substring remaining after
1098−-- dropping characters that fail the predicate @p@ from both the
1216+-- dropping characters that satisfy the predicate @p@ from both the
10991217-- beginning and end of @t@. Subject to fusion.
11001218dropAround :: (Char -> Bool) -> Text -> Text
11011219dropAround p = dropWhile p . dropWhileEnd p
@@ -1130,13 +1248,8 @@ splitAt :: Int -> Text -> (Text, Text)
11301248splitAt n t@(Text arr off len)
11311249 | n <= 0 = (empty, t)
11321250 | n >= len = (t, empty)
1133− | otherwise = (Text arr off k, Text arr (off+k) (len-k))
1134− where k = loop 0 0
1135− loop !i !cnt
1136− | i >= len || cnt >= n = i
1137− | otherwise = loop (i+d) (cnt+1)
1138− where d = iter_ t i
1139−{-# INLINE splitAt #-}
1251+ | otherwise = let k = iterN n t
1252+ in (text arr off k, text arr (off+k) (len-k))
11401253
11411254-- | /O(n)/ 'span', applied to a predicate @p@ and text @t@, returns
11421255-- a pair whose first element is the longest prefix (possibly empty)
@@ -1195,9 +1308,9 @@ tails t | null t = [empty]
11951308-- copies to create substrings; they just construct new 'Text's that
11961309-- are slices of the original.
11971310
1198−-- | /O(m+n)/ Break a 'Text' into pieces separated by the first
1199−-- 'Text' argument, consuming the delimiter. An empty delimiter is
1200−-- invalid, and will cause an error to be raised.
1311+-- | /O(m+n)/ Break a 'Text' into pieces separated by the first 'Text'
1312+-- argument (which cannot be empty), consuming the delimiter. An empty
1313+-- delimiter is invalid, and will cause an error to be raised.
12011314--
12021315-- Examples:
12031316--
@@ -1210,16 +1323,23 @@ tails t | null t = [empty]
12101323-- > intercalate s . splitOn s == id
12111324-- > splitOn (singleton c) == split (==c)
12121325--
1326+-- (Note: the string @s@ to split on above cannot be empty.)
1327+--
12131328-- In (unlikely) bad cases, this function's time complexity degrades
12141329-- towards /O(n*m)/.
1215−splitOn :: Text -> Text -> [Text]
1330+splitOn :: Text
1331+ -- ^ String to split on. If this string is empty, an error
1332+ -- will occur.
1333+ -> Text
1334+ -- ^ Input text.
1335+ -> [Text]
12161336splitOn pat@(Text _ _ l) src@(Text arr off len)
12171337 | l <= 0 = emptyError "splitOn"
12181338 | isSingleton pat = split (== unsafeHead pat) src
12191339 | otherwise = go 0 (indices pat src)
12201340 where
1221− go !s (x:xs) = textP arr (s+off) (x-s) : go (x+l) xs
1222− go s _ = [textP arr (s+off) (len-s)]
1341+ go !s (x:xs) = text arr (s+off) (x-s) : go (x+l) xs
1342+ go s _ = [text arr (s+off) (len-s)]
12231343{-# INLINE [1] splitOn #-}
12241344
12251345{-# RULES
@@ -1311,7 +1431,7 @@ breakOn pat src@(Text arr off len)
13111431 | null pat = emptyError "breakOn"
13121432 | otherwise = case indices pat src of
13131433 [] -> (src, empty)
1314− (x:_) -> (textP arr off x, textP arr (off+x) (len-x))
1434+ (x:_) -> (text arr off x, text arr (off+x) (len-x))
13151435{-# INLINE breakOn #-}
13161436
13171437-- | /O(n+m)/ Similar to 'breakOn', but searches from the end of the
@@ -1353,7 +1473,7 @@ breakOnAll pat src@(Text arr off slen)
13531473 | otherwise = L.map step (indices pat src)
13541474 where
13551475 step x = (chunk 0 x, chunk x (slen-x))
1356− chunk !n !l = textP arr (n+off) l
1476+ chunk !n !l = text arr (n+off) l
13571477{-# INLINE breakOnAll #-}
13581478
13591479-------------------------------------------------------------------------------
@@ -1420,7 +1540,7 @@ countChar c t = S.countChar c (stream t)
14201540-- equivalent to a pair of 'unpack' operations.
14211541zip :: Text -> Text -> [(Char,Char)]
14221542zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1423−{-# INLINE [0] zip #-}
1543+{-# INLINE zip #-}
14241544
14251545-- | /O(n)/ 'zipWith' generalises 'zip' by zipping with the function
14261546-- given as the first argument, instead of a tupling function.
@@ -1428,7 +1548,7 @@ zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
14281548zipWith :: (Char -> Char -> Char) -> Text -> Text -> Text
14291549zipWith f t1 t2 = unstream (S.zipWith g (stream t1) (stream t2))
14301550 where g a b = safe (f a b)
1431−{-# INLINE [0] zipWith #-}
1551+{-# INLINE zipWith #-}
14321552
14331553-- | /O(n)/ Breaks a 'Text' up into a list of words, delimited by 'Char's
14341554-- representing white space.
@@ -1553,7 +1673,7 @@ isInfixOf needle haystack
15531673-- > fnordLength _ = -1
15541674stripPrefix :: Text -> Text -> Maybe Text
15551675stripPrefix p@(Text _arr _off plen) t@(Text arr off len)
1556− | p `isPrefixOf` t = Just $! textP arr (off+plen) (len-plen)
1676+ | p `isPrefixOf` t = Just $! text arr (off+plen) (len-plen)
15571677 | otherwise = Nothing
15581678
15591679-- | /O(n)/ Find the longest non-empty common prefix of two strings
@@ -1573,8 +1693,8 @@ commonPrefixes t0@(Text arr0 off0 len0) t1@(Text arr1 off1 len1) = go 0 0
15731693 where
15741694 go !i !j | i < len0 && j < len1 && a == b = go (i+d0) (j+d1)
15751695 | i > 0 = Just (Text arr0 off0 i,
1576− textP arr0 (off0+i) (len0-i),
1577− textP arr1 (off1+j) (len1-j))
1696+ text arr0 (off0+i) (len0-i),
1697+ text arr1 (off1+j) (len1-j))
15781698 | otherwise = Nothing
15791699 where Iter a d0 = iter t0 i
15801700 Iter b d1 = iter t1 j
@@ -1599,7 +1719,7 @@ commonPrefixes t0@(Text arr0 off0 len0) t1@(Text arr1 off1 len1) = go 0 0
15991719-- > quuxLength _ = -1
16001720stripSuffix :: Text -> Text -> Maybe Text
16011721stripSuffix p@(Text _arr _off plen) t@(Text arr off len)
1602− | p `isSuffixOf` t = Just $! textP arr off (len-plen)
1722+ | p `isSuffixOf` t = Just $! text arr off (len-plen)
16031723 | otherwise = Nothing
16041724
16051725-- | Add a list of non-negative numbers. Errors out on overflow.
@@ -1628,6 +1748,7 @@ overflowError fun = P.error $ "Data.Text." ++ fun ++ ": size overflow"
16281748copy :: Text -> Text
16291749copy (Text arr off len) = Text (A.run go) 0 len
16301750 where
1751+ go :: ST s (A.MArray s)
16311752 go = do
16321753 marr <- A.new len
16331754 A.copyI marr 0 arr off len