@@ -1,15 +1,20 @@
1−{-# LANGUAGE BangPatterns, CPP, Rank2Types #-}
1+{-# LANGUAGE BangPatterns, CPP, MagicHash, Rank2Types, UnboxedTuples #-}
2{-# OPTIONS_GHC -fno-warn-orphans #-}
2{-# OPTIONS_GHC -fno-warn-orphans #-}
3+#if __GLASGOW_HASKELL__ >= 702
4+{-# LANGUAGE Trustworthy #-}
6+#if __GLASGOW_HASKELL__ >= 708
7+{-# LANGUAGE TypeFamilies #-}
5-- Module : Data.Text
11-- Module : Data.Text
6−-- Copyright : (c) 2008, 2009 Tom Harper,
7−-- (c) 2009, 2010 Bryan O'Sullivan,
12+-- Copyright : (c) 2009, 2010, 2011, 2012 Bryan O'Sullivan,
8−-- (c) 2009 Duncan Coutts
13+-- (c) 2009 Duncan Coutts,
14+-- (c) 2008, 2009 Tom Harper
10-- License : BSD-style
16-- License : BSD-style
11−-- Maintainer : bos@serpentine.com, rtomharper@googlemail.com,
17+-- Maintainer : bos@serpentine.com
13-- Stability : experimental
18-- Stability : experimental
14-- Portability : GHC
19-- Portability : GHC
@@ -27,8 +32,8 @@
28-- To use an extended and very rich family of functions for working
33-- To use an extended and very rich family of functions for working
29-- with Unicode text (including normalization, regular expressions,
34-- with Unicode text (including normalization, regular expressions,
30−-- non-standard encodings, text breaking, and locales), see the
35+-- non-standard encodings, text breaking, and locales), see
31−-- @text-icu@ package: <http://hackage.haskell.org/package/text-icu>
36+-- <http://hackage.haskell.org/package/text-icu the text-icu package >.
33module Data.Text
38module Data.Text
@@ -76,6 +81,7 @@ module Data.Text
76 , toCaseFold
81 , toCaseFold
80 -- ** Justification
86 -- ** Justification
81 , justifyLeft
87 , justifyLeft
@@ -119,8 +125,11 @@ module Data.Text
120 -- ** Breaking strings
126 -- ** Breaking strings
123 , takeWhile
131 , takeWhile
124 , dropWhile
133 , dropWhile
125 , dropWhileEnd
134 , dropWhileEnd
126 , dropAround
135 , dropAround
@@ -128,10 +137,10 @@ module Data.Text
128 , stripStart
137 , stripStart
129 , stripEnd
138 , stripEnd
130 , splitAt
139 , splitAt
131− , spanBy
140+ , breakOn
133− , breakEnd
143+ , span
136 , groupBy
145 , groupBy
@@ -139,8 +148,8 @@ module Data.Text
140 -- ** Breaking into many substrings
149 -- ** Breaking into many substrings
141 -- $split
150 -- $split
144 , chunksOf
153 , chunksOf
146 -- ** Breaking into lines and words
155 -- ** Breaking into lines and words
@@ -158,100 +167,119 @@ module Data.Text
158 -- ** View patterns
167 -- ** View patterns
159 , stripPrefix
168 , stripPrefix
160 , stripSuffix
169 , stripSuffix
162 -- * Searching
172 -- * Searching
165− , findBy
176+ , partition
168 -- , findSubstring
178 -- , findSubstring
170 -- * Indexing
180 -- * Indexing
171 -- $index
181 -- $index
173 , findIndex
183 , findIndex
176− -- * Zipping and unzipping
186+ -- * Zipping
178 , zipWith
188 , zipWith
180 -- -* Ordered text
190 -- -* Ordered text
181 -- , sort
191 -- , sort
193+ -- * Low level operations
184import Prelude (Char, Bool(..), Int, Maybe(..), String,
198import Prelude (Char, Bool(..), Int, Maybe(..), String,
185 Eq(..), Ord(..), Ordering(..), (++),
199 Eq(..), Ord(..), Ordering(..), (++),
186− Read(..), Show(..),
200+ Read(..),
187− (&&), (||), (+), (-), (.), ($), ($!), (>>), (*),
201+ (&&), (||), (+), (-), (.), ($), ($!), (>>),
188− div, maxBound, not, return, otherwise)
202+ not, return, otherwise, quot)
189#if defined(HAVE_DEEPSEQ)
203#if defined(HAVE_DEEPSEQ)
190−import Control.DeepSeq (NFData)
204+import Control.DeepSeq (NFData(rnf))
192#if defined(ASSERTS)
206#if defined(ASSERTS)
193import Control.Exception (assert)
207import Control.Exception (assert)
195import Data.Char (isSpace)
209import Data.Char (isSpace)
196−import Data.Data (Data(gfoldl, toConstr, gunfold, dataTypeOf))
210+import Data.Data (Data(gfoldl, toConstr, gunfold, dataTypeOf), constrIndex,
197−#if __GLASGOW_HASKELL__ >= 612
211+ Constr, mkConstr, DataType, mkDataType, Fixity(Prefix))
198−import Data.Data (mkNoRepType)
200−import Data.Data (mkNorepType)
202import Control.Monad (foldM)
212import Control.Monad (foldM)
213+import Control.Monad.ST (ST)
203import qualified Data.Text.Array as A
214import qualified Data.Text.Array as A
204import qualified Data.List as L
215import qualified Data.List as L
216+import Data.Binary (Binary(get, put))
205import Data.Monoid (Monoid(..))
217import Data.Monoid (Monoid(..))
218+#if MIN_VERSION_base(4,9,0)
219+import Data.Semigroup (Semigroup(..))
206import Data.String (IsString(..))
221import Data.String (IsString(..))
207−import qualified Data.Text.Fusion as S
222+import qualified Data.Text.Internal.Fusion as S
208−import qualified Data.Text.Fusion.Common as S
223+import qualified Data.Text.Internal.Fusion.Common as S
224+import Data.Text.Encoding (decodeUtf8', encodeUtf8)
209−import Data.Text.Fusion (stream, reverseStream, unstream)
225+import Data.Text.Internal.Fusion (stream, reverseStream, unstream)
210−import Data.Text.Internal (Text(..), empty, firstf, safe, text, textP)
226+import Data.Text.Internal.Private (span_)
227+import Data.Text.Internal (Text(..), empty, firstf, mul, safe, text)
228+import Data.Text.Show (singleton, unpack, unpackCString#)
211import qualified Prelude as P
229import qualified Prelude as P
212import Data.Text.Unsafe (Iter(..), iter, iter_, lengthWord16, reverseIter,
230import Data.Text.Unsafe (Iter(..), iter, iter_, lengthWord16, reverseIter,
213− unsafeHead, unsafeTail)
231+ reverseIter_, unsafeHead, unsafeTail)
214−import Data.Text.UnsafeChar (unsafeChr)
232+import Data.Text.Internal.Unsafe.Char (unsafeChr)
215−import qualified Data.Text.Util as U
233+import qualified Data.Text.Internal.Functions as F
216−import qualified Data.Text.Encoding.Utf16 as U16
234+import qualified Data.Text.Internal.Encoding.Utf16 as U16
217−import Data.Text.Search (indices)
235+import Data.Text.Internal.Search (indices)
218#if defined(__HADDOCK__)
236#if defined(__HADDOCK__)
219import Data.ByteString (ByteString)
237import Data.ByteString (ByteString)
220import qualified Data.Text.Lazy as L
238import qualified Data.Text.Lazy as L
221import Data.Int (Int64)
239import Data.Int (Int64)
241+import GHC.Base (eqInt, neInt, gtInt, geInt, ltInt, leInt)
242+#if __GLASGOW_HASKELL__ >= 708
243+import qualified GHC.Exts as Exts
245+#if MIN_VERSION_base(4,7,0)
246+import Text.Printf (PrintfArg, formatArg, formatString)
224-- $strict
249-- $strict
226-- This package provides both strict and lazy 'Text' types. The
251-- This package provides both strict and lazy 'Text' types. The
227−-- strict type is provided by the 'Data.Text' package, while the lazy
252+-- strict type is provided by the "Data.Text" module, while the lazy
228−-- type is provided by the 'Data.Text.Lazy' package. Internally, the
253+-- type is provided by the "Data.Text.Lazy" module. Internally, the
229-- lazy @Text@ type consists of a list of strict chunks.
254-- lazy @Text@ type consists of a list of strict chunks.
231-- The strict 'Text' type requires that an entire string fit into
256-- The strict 'Text' type requires that an entire string fit into
232−-- memory at once. The lazy @Text@ type is capable of streaming
257+-- memory at once. The lazy 'Data.Text.Lazy.Text' type is capable of
233−-- strings that are larger than memory using a small memory footprint.
258+-- streaming strings that are larger than memory using a small memory
234−-- In many cases, the overhead of chunked streaming makes the lazy
259+-- footprint. In many cases, the overhead of chunked streaming makes
235−-- @Text@ type slower than its strict counterpart, but this is not
260+-- the lazy 'Data.Text.Lazy.Text' type slower than its strict
236−-- always the case. Sometimes, the time complexity of a function in
261+-- counterpart, but this is not always the case. Sometimes, the time
237−-- one module may be different from the other, due to their differing
262+-- complexity of a function in one module may be different from the
238−-- internal structures.
263+-- other, due to their differing internal structures.
240-- Each module provides an almost identical API, with the main
265-- Each module provides an almost identical API, with the main
241-- difference being that the strict module uses 'Int' values for
266-- difference being that the strict module uses 'Int' values for
242−-- lengths and counts, while the lazy module uses 'Int64' lengths.
267+-- lengths and counts, while the lazy module uses 'Data.Int.Int64'
244-- $replacement
270-- $replacement
246-- A 'Text' value is a sequence of Unicode scalar values, as defined
272-- A 'Text' value is a sequence of Unicode scalar values, as defined
247−-- in §3.9, definition D76 of the Unicode 5.2 standard:
274+-- <http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=35 §3.9, definition D76 of the Unicode 5.2 standard >.
248−-- <http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=35>. As
249−-- such, a 'Text' cannot contain values in the range U+D800 to U+DFFF
275+-- As such, a 'Text' cannot contain values in the range U+D800 to
250−-- inclusive. Haskell implementations admit all Unicode code points
276+-- U+DFFF inclusive. Haskell implementations admit all Unicode code
278+-- (<http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=13 §3.4, definition D10 >)
251−-- (§3.4, definition D10) as 'Char' values, including code points
279+-- as 'Char' values, including code points from this invalid range.
252−-- from this invalid range. This means that there are some 'Char'
280+-- This means that there are some 'Char' values that are not valid
253−-- values that are not valid Unicode scalar values, and the functions
281+-- Unicode scalar values, and the functions in this module must handle
254−-- in this module must handle those cases.
282+-- those cases.
256-- Within this module, many functions construct a 'Text' from one or
284-- Within this module, many functions construct a 'Text' from one or
257-- more 'Char' values. Those functions will substitute 'Char' values
285-- more 'Char' values. Those functions will substitute 'Char' values
@@ -265,8 +293,8 @@ import Data.Int (Int64)
265-- range U+D800 through U+DFFF are used by UTF-16 to denote surrogate
293-- range U+D800 through U+DFFF are used by UTF-16 to denote surrogate
266-- code points, and so cannot be represented. The functions replace
294-- code points, and so cannot be represented. The functions replace
267-- invalid scalar values, instead of dropping them, as a security
295-- invalid scalar values, instead of dropping them, as a security
268−-- measure. For details, see Unicode Technical Report 36, §3.5:
296+-- measure. For details, see
269−-- <http://unicode.org/reports/tr36/#Deletion_of_Noncharacters>)
297+-- <http://unicode.org/reports/tr36/#Deletion_of_Noncharacters Unicode Technical Report 36, §3.5 >.)
271-- $fusion
299-- $fusion
@@ -284,11 +312,11 @@ import Data.Int (Int64)
284-- > countChars = T.length . T.toUpper . E.decodeUtf8
312-- > countChars = T.length . T.toUpper . E.decodeUtf8
286-- From the type signatures involved, this looks like it should
314-- From the type signatures involved, this looks like it should
287−-- allocate one 'ByteString' value, and two 'Text' values. However,
315+-- allocate one 'Data.ByteString.ByteString' value, and two 'Text'
288−-- when a module is compiled with optimisation enabled under GHC, the
316+-- values. However, when a module is compiled with optimisation
289−-- two intermediate 'Text' values will be optimised away, and the
317+-- enabled under GHC, the two intermediate 'Text' values will be
290−-- function will be compiled down to a single loop over the source
318+-- optimised away, and the function will be compiled down to a single
291−-- 'ByteString'.
319+-- loop over the source 'Data.ByteString.ByteString'.
293-- Functions that can be fused by the compiler are documented with the
321-- Functions that can be fused by the compiler are documented with the
294-- phrase \"Subject to fusion\".
322-- phrase \"Subject to fusion\".
@@ -302,46 +330,83 @@ instance Eq Text where
302instance Ord Text where
330instance Ord Text where
303 compare = compareText
331 compare = compareText
305−instance Show Text where
306− showsPrec p ps r = showsPrec p (unpack ps) r
308instance Read Text where
333instance Read Text where
309 readsPrec p str = [(pack x,y) | (x,y) <- readsPrec p str]
334 readsPrec p str = [(pack x,y) | (x,y) <- readsPrec p str]
336+#if MIN_VERSION_base(4,9,0)
337+-- Semigroup orphan instances for older GHCs are provided by
338+-- 'semigroups` package
340+instance Semigroup Text where
311instance Monoid Text where
344instance Monoid Text where
312 mempty = empty
345 mempty = empty
346+#if MIN_VERSION_base(4,9,0)
347+ mappend = (<>) -- future-proof definition
313 mappend = append
349 mappend = append
314 mconcat = concat
351 mconcat = concat
316instance IsString Text where
353instance IsString Text where
317 fromString = pack
354 fromString = pack
356+#if __GLASGOW_HASKELL__ >= 708
357+instance Exts.IsList Text where
358+ type Item Text = Char
319#if defined(HAVE_DEEPSEQ)
363#if defined(HAVE_DEEPSEQ)
320−instance NFData Text
364+instance NFData Text where rnf !_ = ()
367+instance Binary Text where
368+ put t = put (encodeUtf8 t)
371+ case decodeUtf8' bs of
372+ P.Left exn -> P.fail (P.show exn)
373+ P.Right a -> P.return a
323−-- This instance preserves data abstraction at the cost of inefficiency.
375+-- | This instance preserves data abstraction at the cost of inefficiency.
324-- We omit reflection services for the sake of data abstraction.
376-- We omit reflection services for the sake of data abstraction.
326−-- This instance was created by copying the behavior of Data.Set and
327−-- Data.Map. If you feel a mistake has been made, please feel free to
328−-- submit improvements.
330−-- Original discussion is archived here:
378+-- This instance was created by copying the updated behavior of
331−
379+-- @"Data.Set".@'Data.Set.Set' and @"Data.Map".@'Data.Map.Map'. If you
332−-- "could we get a Data instance for Data.Text.Text?"
380+-- feel a mistake has been made, please feel free to submit
383+-- The original discussion is archived here:
333−-- http://groups.google.com/group/haskell-cafe/browse_thread/thread/b5bbb1b28a7e525d/0639d46852575b93
384+-- <http://groups.google.com/group/haskell-cafe/browse_thread/thread/b5bbb1b28a7e525d/0639d46852575b93 could we get a Data instance for Data.Text.Text? >
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, ... >
335instance Data Text where
390instance Data Text where
336 gfoldl f z txt = z pack `f` (unpack txt)
391 gfoldl f z txt = z pack `f` (unpack txt)
337− toConstr _ = P.error "Data.Text.Text.toConstr"
392+ toConstr _ = packConstr
393+ gunfold k z c = case constrIndex c of
338− gunfold _ _ = P.error "Data.Text.Text.gunfold"
395+ _ -> P.error "gunfold"
339−#if __GLASGOW_HASKELL__ >= 612
396+ dataTypeOf _ = textDataType
340− dataTypeOf _ = mkNoRepType "Data.Text.Text"
397+
341−#else
398+#if MIN_VERSION_base(4,7,0)
342− dataTypeOf _ = mkNorepType "Data.Text.Text"
399+-- | Only defined for @base-4.7.0.0@ and later
400+instance PrintfArg Text where
401+ formatArg txt = formatString $ unpack txt
405+packConstr = mkConstr textDataType "pack" [] Prefix
407+textDataType :: DataType
408+textDataType = mkDataType "Data.Text.Text" [packConstr]
345-- | /O(n)/ Compare two 'Text' values lexicographically.
410-- | /O(n)/ Compare two 'Text' values lexicographically.
346compareText :: Text -> Text -> Ordering
411compareText :: Text -> Text -> Ordering
347compareText ta@(Text _arrA _offA lenA) tb@(Text _arrB _offB lenB)
412compareText ta@(Text _arrA _offA lenA) tb@(Text _arrB _offB lenB)
@@ -362,20 +427,9 @@ compareText ta@(Text _arrA _offA lenA) tb@(Text _arrB _offB lenB)
362-- | /O(n)/ Convert a 'String' into a 'Text'. Subject to
427-- | /O(n)/ Convert a 'String' into a 'Text'. Subject to
363-- fusion. Performs replacement on invalid scalar values.
428-- fusion. Performs replacement on invalid scalar values.
364pack :: String -> Text
429pack :: String -> Text
365−pack = unstream . S.streamList . L.map safe
430+pack = unstream . S.map safe . S.streamList
366{-# INLINE [1] pack #-}
431{-# INLINE [1] pack #-}
368−-- | /O(n)/ Convert a Text into a String. Subject to fusion.
369−unpack :: Text -> String
370−unpack = S.unstreamList . stream
371−{-# INLINE [1] unpack #-}
373−-- | /O(1)/ Convert a character into a Text. Subject to fusion.
374−-- Performs replacement on invalid scalar values.
375−singleton :: Char -> Text
376−singleton = unstream . S.singleton . safe
377−{-# INLINE [1] singleton #-}
379-- -----------------------------------------------------------------------------
433-- -----------------------------------------------------------------------------
380-- * Basic functions
434-- * Basic functions
@@ -387,6 +441,8 @@ cons :: Char -> Text -> Text
387cons c t = unstream (S.cons (safe c) (stream t))
441cons c t = unstream (S.cons (safe c) (stream t))
388{-# INLINE cons #-}
442{-# INLINE cons #-}
390-- | /O(n)/ Adds a character to the end of a 'Text'. This copies the
446-- | /O(n)/ Adds a character to the end of a 'Text'. This copies the
391-- entire array in the process, unless fused. Subject to fusion.
447-- entire array in the process, unless fused. Subject to fusion.
392-- Performs replacement on invalid scalar values.
448-- Performs replacement on invalid scalar values.
@@ -404,12 +460,13 @@ append a@(Text arr1 off1 len1) b@(Text arr2 off2 len2)
404 | otherwise = overflowError "append"
460 | otherwise = overflowError "append"
406 len = len1+len2
462 len = len1+len2
463+ x :: ST s (A.MArray s)
408 arr <- A.new len
465 arr <- A.new len
409 A.copyI arr 0 arr1 off1 len1
466 A.copyI arr 0 arr1 off1 len1
410 A.copyI arr len1 arr2 off2 len
467 A.copyI arr len1 arr2 off2 len
411 return arr
468 return arr
412−{-# INLINE append #-}
469+{-# NOINLINE append #-}
415"TEXT append -> fused" [~1] forall t1 t2.
472"TEXT append -> fused" [~1] forall t1 t2.
@@ -429,8 +486,8 @@ head t = S.head (stream t)
429uncons :: Text -> Maybe (Char, Text)
486uncons :: Text -> Maybe (Char, Text)
430uncons t@(Text arr off len)
487uncons t@(Text arr off len)
431 | len <= 0 = Nothing
488 | len <= 0 = Nothing
489+ | otherwise = Just $ let !(Iter c d) = iter t 0
432− | otherwise = Just (c, textP arr (off+d) (len-d))
490+ in (c, text arr (off+d) (len-d))
433− where Iter c d = iter t 0
434{-# INLINE [1] uncons #-}
491{-# INLINE [1] uncons #-}
436-- | Lifted from Control.Arrow and specialized.
493-- | Lifted from Control.Arrow and specialized.
@@ -460,7 +517,7 @@ last (Text arr off len)
460tail :: Text -> Text
517tail :: Text -> Text
461tail t@(Text arr off len)
518tail t@(Text arr off len)
462 | len <= 0 = emptyError "tail"
519 | len <= 0 = emptyError "tail"
463− | otherwise = textP arr (off+d) (len-d)
520+ | otherwise = text arr (off+d) (len-d)
464 where d = iter_ t 0
521 where d = iter_ t 0
465{-# INLINE [1] tail #-}
522{-# INLINE [1] tail #-}
@@ -475,8 +532,8 @@ tail t@(Text arr off len)
475-- be non-empty. Subject to fusion.
532-- be non-empty. Subject to fusion.
476init :: Text -> Text
533init :: Text -> Text
477init (Text arr off len) | len <= 0 = emptyError "init"
534init (Text arr off len) | len <= 0 = emptyError "init"
478− | n >= 0xDC00 && n <= 0xDFFF = textP arr off (len-2)
535+ | n >= 0xDC00 && n <= 0xDFFF = text arr off (len-2)
479− | otherwise = textP arr off (len-1)
536+ | otherwise = text arr off (len-1)
481 n = A.unsafeIndex arr (off+len-1)
538 n = A.unsafeIndex arr (off+len-1)
482{-# INLINE [1] init #-}
539{-# INLINE [1] init #-}
@@ -515,7 +572,9 @@ isSingleton = S.isSingleton . stream
515-- Subject to fusion.
572-- Subject to fusion.
516length :: Text -> Int
573length :: Text -> Int
517length t = S.length (stream t)
574length t = S.length (stream t)
518−{-# 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.
520-- | /O(n)/ Compare the count of characters in a 'Text' to a number.
579-- | /O(n)/ Compare the count of characters in a 'Text' to a number.
521-- Subject to fusion.
580-- Subject to fusion.
@@ -534,32 +593,32 @@ compareLength t n = S.compareLengthI (stream t) n
536"TEXT ==N/length -> compareLength/==EQ" [~1] forall t n.
595"TEXT ==N/length -> compareLength/==EQ" [~1] forall t n.
537− (==) (length t) n = compareLength t n == EQ
596+ eqInt (length t) n = compareLength t n == EQ
541"TEXT /=N/length -> compareLength//=EQ" [~1] forall t n.
600"TEXT /=N/length -> compareLength//=EQ" [~1] forall t n.
542− (/=) (length t) n = compareLength t n /= EQ
601+ neInt (length t) n = compareLength t n /= EQ
546"TEXT <N/length -> compareLength/==LT" [~1] forall t n.
605"TEXT <N/length -> compareLength/==LT" [~1] forall t n.
547− (<) (length t) n = compareLength t n == LT
606+ ltInt (length t) n = compareLength t n == LT
551"TEXT <=N/length -> compareLength//=GT" [~1] forall t n.
610"TEXT <=N/length -> compareLength//=GT" [~1] forall t n.
552− (<=) (length t) n = compareLength t n /= GT
611+ leInt (length t) n = compareLength t n /= GT
556"TEXT >N/length -> compareLength/==GT" [~1] forall t n.
615"TEXT >N/length -> compareLength/==GT" [~1] forall t n.
557− (>) (length t) n = compareLength t n == GT
616+ gtInt (length t) n = compareLength t n == GT
561"TEXT >=N/length -> compareLength//=LT" [~1] forall t n.
620"TEXT >=N/length -> compareLength//=LT" [~1] forall t n.
562− (>=) (length t) n = compareLength t n /= LT
621+ geInt (length t) n = compareLength t n /= LT
565-- -----------------------------------------------------------------------------
624-- -----------------------------------------------------------------------------
@@ -575,7 +634,7 @@ map f t = unstream (S.map (safe . f) (stream t))
575-- 'Text's and concatenates the list after interspersing the first
634-- 'Text's and concatenates the list after interspersing the first
576-- argument between each element of the list.
635-- argument between each element of the list.
577intercalate :: Text -> [Text] -> Text
636intercalate :: Text -> [Text] -> Text
578−intercalate t = concat . (U.intersperse t)
637+intercalate t = concat . (F.intersperse t)
579{-# INLINE intercalate #-}
638{-# INLINE intercalate #-}
581-- | /O(n)/ The 'intersperse' function takes a character and places it
640-- | /O(n)/ The 'intersperse' function takes a character and places it
@@ -590,16 +649,59 @@ reverse :: Text -> Text
590reverse t = S.reverse (stream t)
649reverse t = S.reverse (stream t)
591{-# INLINE reverse #-}
650{-# INLINE reverse #-}
593−-- | /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@.
655+-- This function behaves as though it was defined as follows:
658+-- replace needle replacement haystack =
659+-- 'intercalate' replacement ('splitOn' needle haystack)
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:
666+-- > replace "oo" "foo" "oo" == "foo"
668+-- In cases where several instances of @needle@ overlap, only the
669+-- first one will be replaced:
671+-- > replace "ofo" "bar" "ofofo" == "barfo"
595-- In (unlikely) bad cases, this function's time complexity degrades
673-- In (unlikely) bad cases, this function's time complexity degrades
596-- towards /O(n*m)/.
674-- towards /O(n*m)/.
597−replace :: Text -- ^ Text to search for
675+replace :: Text
598− -> Text -- ^ Replacement text
676+ -- ^ @needle@ to search for. If this string is empty, an
677+ -- error will occur.
679+ -- ^ @replacement@ to replace @needle@ with.
599− -> Text -- ^ Input text
680+ -> Text
681+ -- ^ @haystack@ in which to search.
601−replace s d = intercalate d . split s
683+replace needle@(Text _ _ neeLen)
602−{-# INLINE replace #-}
684+ (Text repArr repOff repLen)
685+ haystack@(Text hayArr hayOff hayLen)
686+ | neeLen == 0 = emptyError "replace"
687+ | L.null ixs = haystack
688+ | len > 0 = Text (A.run x) 0 len
691+ ixs = indices needle haystack
692+ len = hayLen - (neeLen - repLen) `mul` L.length ixs
693+ x :: ST s (A.MArray s)
696+ let loop (i:is) o d = do
699+ A.copyI marr d hayArr (hayOff+o) d0
700+ A.copyI marr d0 repArr repOff d1
701+ loop is (i + neeLen) d1
702+ loop [] o d = A.copyI marr d hayArr (hayOff+o) len
604-- ----------------------------------------------------------------------------
706-- ----------------------------------------------------------------------------
605-- ** Case conversions (folds)
707-- ** Case conversions (folds)
@@ -618,12 +720,13 @@ replace s d = intercalate d . split s
618-- /Note/: In some languages, case conversion is a locale- and
720-- /Note/: In some languages, case conversion is a locale- and
619-- context-dependent operation. The case conversion functions in this
721-- context-dependent operation. The case conversion functions in this
620-- module are /not/ locale sensitive. Programs that require locale
722-- module are /not/ locale sensitive. Programs that require locale
621−-- sensitivity should use appropriate versions of the case mapping
723+-- sensitivity should use appropriate versions of the
622−-- functions from the @text-icu@ package.
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 >.
624−-- | /O(n)/ Convert a string to folded case. This function is mainly
726+-- | /O(n)/ Convert a string to folded case. Subject to fusion.
625−-- useful for performing caseless (also known as case insensitive)
728+-- This function is mainly useful for performing caseless (also known
626−-- string comparisons.
729+-- as case insensitive) string comparisons.
628-- A string @x@ is a caseless match for a string @y@ if and only if:
731-- A string @x@ is a caseless match for a string @y@ if and only if:
@@ -638,25 +741,51 @@ replace s d = intercalate d . split s
638-- instead of itself.
741-- instead of itself.
639toCaseFold :: Text -> Text
742toCaseFold :: Text -> Text
640toCaseFold t = unstream (S.toCaseFold (stream t))
743toCaseFold t = unstream (S.toCaseFold (stream t))
641−{-# INLINE [0] toCaseFold #-}
744+{-# INLINE toCaseFold #-}
643-- | /O(n)/ Convert a string to lower case, using simple case
746-- | /O(n)/ Convert a string to lower case, using simple case
747+-- conversion. Subject to fusion.
644−-- conversion. The result string may be longer than the input string.
749+-- The result string may be longer than the input string. For
645−-- For instance, \"İ\" (Latin capital letter I with dot above,
750+-- instance, \"İ\" (Latin capital letter I with dot above,
646−-- U+0130) maps to the sequence \"i\" (Latin small letter i, U+0069) followed
751+-- U+0130) maps to the sequence \"i\" (Latin small letter i, U+0069)
647−-- by \" ̇\" (combining dot above, U+0307).
752+-- followed by \" ̇\" (combining dot above, U+0307).
648toLower :: Text -> Text
753toLower :: Text -> Text
649toLower t = unstream (S.toLower (stream t))
754toLower t = unstream (S.toLower (stream t))
650{-# INLINE toLower #-}
755{-# INLINE toLower #-}
652-- | /O(n)/ Convert a string to upper case, using simple case
757-- | /O(n)/ Convert a string to upper case, using simple case
758+-- conversion. Subject to fusion.
653−-- conversion. The result string may be longer than the input string.
760+-- The result string may be longer than the input string. For
654−-- For instance, the German \"ß\" (eszett, U+00DF) maps to the
761+-- instance, the German \"ß\" (eszett, U+00DF) maps to the
655-- two-letter sequence \"SS\".
762-- two-letter sequence \"SS\".
656toUpper :: Text -> Text
763toUpper :: Text -> Text
657toUpper t = unstream (S.toUpper (stream t))
764toUpper t = unstream (S.toUpper (stream t))
658{-# INLINE toUpper #-}
765{-# INLINE toUpper #-}
767+-- | /O(n)/ Convert a string to title case, using simple case
768+-- conversion. Subject to fusion.
770+-- The first letter of the input is converted to title case, as is
771+-- every subsequent letter that immediately follows a non-letter.
772+-- Every letter that immediately follows another letter is converted
775+-- The result string may be longer than the input string. For example,
776+-- the Latin small ligature fl (U+FB02) is converted to the
777+-- sequence Latin capital letter F (U+0046) followed by Latin small
778+-- letter l (U+006C).
780+-- /Note/: this function does not take language or culture specific
781+-- rules into account. For instance, in English, different style
782+-- guides disagree on whether the book name \"The Hill of the Red
783+-- Fox\" is correctly title cased—but this function will
784+-- capitalize /every/ word.
785+toTitle :: Text -> Text
786+toTitle t = unstream (S.toTitle (stream t))
787+{-# INLINE toTitle #-}
660-- | /O(n)/ Left-justify a string to the given length, using the
789-- | /O(n)/ Left-justify a string to the given length, using the
661-- specified fill character on the right. Subject to fusion.
790-- specified fill character on the right. Subject to fusion.
662-- Performs replacement on invalid scalar values.
791-- Performs replacement on invalid scalar values.
@@ -707,7 +836,7 @@ center k c t
707 | otherwise = replicateChar l c `append` t `append` replicateChar r c
836 | otherwise = replicateChar l c `append` t `append` replicateChar r c
708 where len = length t
837 where len = length t
709 d = k - len
838 d = k - len
710− r = d `div` 2
839+ r = d `quot` 2
711 l = d - r
840 l = d - r
712{-# INLINE center #-}
841{-# INLINE center #-}
@@ -754,7 +883,7 @@ foldr f z t = S.foldr f z (stream t)
754{-# INLINE foldr #-}
883{-# INLINE foldr #-}
756-- | /O(n)/ A variant of 'foldr' that has no starting value argument,
885-- | /O(n)/ A variant of 'foldr' that has no starting value argument,
757−-- and thust must be applied to a non-empty 'Text'. Subject to
886+-- and thus must be applied to a non-empty 'Text'. Subject to
758-- fusion.
887-- fusion.
759foldr1 :: (Char -> Char -> Char) -> Text -> Char
888foldr1 :: (Char -> Char -> Char) -> Text -> Char
760foldr1 f t = S.foldr1 f (stream t)
889foldr1 f t = S.foldr1 f (stream t)
@@ -772,6 +901,7 @@ concat ts = case ts' of
773 ts' = L.filter (not . null) ts
902 ts' = L.filter (not . null) ts
774 len = sumP "concat" $ L.map lengthWord16 ts'
903 len = sumP "concat" $ L.map lengthWord16 ts'
904+ go :: ST s (A.MArray s)
776 arr <- A.new len
906 arr <- A.new len
777 let step i (Text a o l) =
907 let step i (Text a o l) =
@@ -785,13 +915,13 @@ concatMap f = concat . foldr ((:) . f) []
785{-# INLINE concatMap #-}
915{-# INLINE concatMap #-}
787-- | /O(n)/ 'any' @p@ @t@ determines whether any character in the
917-- | /O(n)/ 'any' @p@ @t@ determines whether any character in the
788−-- 'Text' @t@ satisifes the predicate @p@. Subject to fusion.
918+-- 'Text' @t@ satisfies the predicate @p@. Subject to fusion.
789any :: (Char -> Bool) -> Text -> Bool
919any :: (Char -> Bool) -> Text -> Bool
790any p t = S.any p (stream t)
920any p t = S.any p (stream t)
791{-# INLINE any #-}
921{-# INLINE any #-}
793-- | /O(n)/ 'all' @p@ @t@ determines whether all characters in the
923-- | /O(n)/ 'all' @p@ @t@ determines whether all characters in the
794−-- 'Text' @t@ satisify the predicate @p@. Subject to fusion.
924+-- 'Text' @t@ satisfy the predicate @p@. Subject to fusion.
795all :: (Char -> Bool) -> Text -> Bool
925all :: (Char -> Bool) -> Text -> Bool
796all p t = S.all p (stream t)
926all p t = S.all p (stream t)
797{-# INLINE all #-}
927{-# INLINE all #-}
@@ -879,13 +1009,13 @@ mapAccumR f z0 = second reverse . S.mapAccumL g z0 . reverseStream
879-- @t@ repeated @n@ times.
1009-- @t@ repeated @n@ times.
880replicate :: Int -> Text -> Text
1010replicate :: Int -> Text -> Text
881replicate n t@(Text a o l)
1011replicate n t@(Text a o l)
882− | n <= 0 || l <= 0 = empty
1012+ | n <= 0 || l <= 0 = empty
883− | n == 1 = t
1013+ | n == 1 = t
884− | isSingleton t = replicateChar n (unsafeHead t)
1014+ | isSingleton t = replicateChar n (unsafeHead t)
885− | n <= maxBound `div` l = Text (A.run x) 0 len
1015+ | otherwise = Text (A.run x) 0 len
886− | otherwise = overflowError "replicate"
888− len = l * n
1017+ len = l `mul` n
1018+ x :: ST s (A.MArray s)
890 arr <- A.new len
1020 arr <- A.new len
891 let loop !d !i | i >= n = return arr
1021 let loop !d !i | i >= n = return arr
@@ -936,14 +1066,16 @@ take :: Int -> Text -> Text
936take n t@(Text arr off len)
1066take n t@(Text arr off len)
937 | n <= 0 = empty
1067 | n <= 0 = empty
938 | n >= len = t
1068 | n >= len = t
939− | otherwise = Text arr off (loop 0 0)
1069+ | otherwise = text arr off (iterN n t)
942− | i >= len || cnt >= n = i
943− | otherwise = loop (i+d) (cnt+1)
945{-# INLINE [1] take #-}
1070{-# INLINE [1] take #-}
1072+iterN :: Int -> Text -> Int
1073+iterN n t@(Text _arr _off len) = loop 0 0
1075+ | i >= len || cnt >= n = i
1076+ | otherwise = loop (i+d) (cnt+1)
1077+ where d = iter_ t i
947{-# RULES
1079{-# RULES
948"TEXT take -> fused" [~1] forall n t.
1080"TEXT take -> fused" [~1] forall n t.
949 take n t = unstream (S.take n (stream t))
1081 take n t = unstream (S.take n (stream t))
@@ -951,6 +1083,27 @@ take n t@(Text arr off len)
951 unstream (S.take n (stream t)) = take n t
1083 unstream (S.take n (stream t)) = take n t
1086+-- | /O(n)/ 'takeEnd' @n@ @t@ returns the suffix remaining after
1087+-- taking @n@ characters from the end of @t@.
1091+-- > takeEnd 3 "foobar" == "bar"
1092+takeEnd :: Int -> Text -> Text
1093+takeEnd n t@(Text arr off len)
1096+ | otherwise = text arr (off+i) (len-i)
1097+ where i = iterNEnd n t
1099+iterNEnd :: Int -> Text -> Int
1100+iterNEnd n t@(Text _arr _off len) = loop (len-1) n
1104+ | otherwise = loop (i+d) (m-1)
1105+ where d = reverseIter_ t i
954-- | /O(n)/ 'drop' @n@, applied to a 'Text', returns the suffix of the
1107-- | /O(n)/ 'drop' @n@, applied to a 'Text', returns the suffix of the
955-- 'Text' after the first @n@ characters, or the empty 'Text' if @n@
1108-- 'Text' after the first @n@ characters, or the empty 'Text' if @n@
956-- is greater than the length of the 'Text'. Subject to fusion.
1109-- is greater than the length of the 'Text'. Subject to fusion.
@@ -958,11 +1111,8 @@ drop :: Int -> Text -> Text
958drop n t@(Text arr off len)
1111drop n t@(Text arr off len)
959 | n <= 0 = t
1112 | n <= 0 = t
960 | n >= len = empty
1113 | n >= len = empty
961− | otherwise = loop 0 0
1114+ | otherwise = text arr (off+i) (len-i)
963− | i >= len || cnt >= n = Text arr (off+i) (len-i)
964− | otherwise = loop (i+d) (cnt+1)
965− where d = iter_ t i
1115+ where i = iterN n t
966{-# INLINE [1] drop #-}
1116{-# INLINE [1] drop #-}
968{-# RULES
1118{-# RULES
@@ -972,6 +1122,18 @@ drop n t@(Text arr off len)
972 unstream (S.drop n (stream t)) = drop n t
1122 unstream (S.drop n (stream t)) = drop n t
1125+-- | /O(n)/ 'dropEnd' @n@ @t@ returns the prefix remaining after
1126+-- dropping @n@ characters from the end of @t@.
1130+-- > dropEnd 3 "foobar" == "foo"
1131+dropEnd :: Int -> Text -> Text
1132+dropEnd n t@(Text arr off len)
1135+ | otherwise = text arr off (iterNEnd n t)
975-- | /O(n)/ 'takeWhile', applied to a predicate @p@ and a 'Text',
1137-- | /O(n)/ 'takeWhile', applied to a predicate @p@ and a 'Text',
976-- returns the longest prefix (possibly empty) of elements that
1138-- returns the longest prefix (possibly empty) of elements that
977-- satisfy @p@. Subject to fusion.
1139-- satisfy @p@. Subject to fusion.
@@ -979,7 +1141,7 @@ takeWhile :: (Char -> Bool) -> Text -> Text
979takeWhile p t@(Text arr off len) = loop 0
1141takeWhile p t@(Text arr off len) = loop 0
980 where loop !i | i >= len = t
1142 where loop !i | i >= len = t
981 | p c = loop (i+d)
1143 | p c = loop (i+d)
982− | otherwise = textP arr off i
1144+ | otherwise = text arr off i
983 where Iter c d = iter t i
1145 where Iter c d = iter t i
984{-# INLINE [1] takeWhile #-}
1146{-# INLINE [1] takeWhile #-}
@@ -990,6 +1152,27 @@ takeWhile p t@(Text arr off len) = loop 0
990 unstream (S.takeWhile p (stream t)) = takeWhile p t
1152 unstream (S.takeWhile p (stream t)) = takeWhile p t
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.
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 #-}
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
993-- | /O(n)/ 'dropWhile' @p@ @t@ returns the suffix remaining after
1176-- | /O(n)/ 'dropWhile' @p@ @t@ returns the suffix remaining after
994-- 'takeWhile' @p@ @t@. Subject to fusion.
1177-- 'takeWhile' @p@ @t@. Subject to fusion.
995dropWhile :: (Char -> Bool) -> Text -> Text
1178dropWhile :: (Char -> Bool) -> Text -> Text
@@ -1008,8 +1191,9 @@ dropWhile p t@(Text arr off len) = loop 0 0
1010-- | /O(n)/ 'dropWhileEnd' @p@ @t@ returns the prefix remaining after
1193-- | /O(n)/ 'dropWhileEnd' @p@ @t@ returns the prefix remaining after
1011−-- dropping characters that fail the predicate @p@ from the end of
1194+-- dropping characters that satisfy the predicate @p@ from the end of
1012-- @t@. Subject to fusion.
1195-- @t@. Subject to fusion.
1013-- Examples:
1197-- Examples:
1015-- > dropWhileEnd (=='.') "foo..." == "foo"
1199-- > dropWhileEnd (=='.') "foo..." == "foo"
@@ -1029,7 +1213,7 @@ dropWhileEnd p t@(Text arr off len) = loop (len-1) len
1031-- | /O(n)/ 'dropAround' @p@ @t@ returns the substring remaining after
1215-- | /O(n)/ 'dropAround' @p@ @t@ returns the substring remaining after
1032−-- dropping characters that fail the predicate @p@ from both the
1216+-- dropping characters that satisfy the predicate @p@ from both the
1033-- beginning and end of @t@. Subject to fusion.
1217-- beginning and end of @t@. Subject to fusion.
1034dropAround :: (Char -> Bool) -> Text -> Text
1218dropAround :: (Char -> Bool) -> Text -> Text
1035dropAround p = dropWhile p . dropWhileEnd p
1219dropAround p = dropWhile p . dropWhileEnd p
@@ -1064,31 +1248,23 @@ splitAt :: Int -> Text -> (Text, Text)
1064splitAt n t@(Text arr off len)
1248splitAt n t@(Text arr off len)
1065 | n <= 0 = (empty, t)
1249 | n <= 0 = (empty, t)
1066 | n >= len = (t, empty)
1250 | n >= len = (t, empty)
1067− | otherwise = (Text arr off k, Text arr (off+k) (len-k))
1251+ | otherwise = let k = iterN n t
1068− where k = loop 0 0
1252+ in (text arr off k, text arr (off+k) (len-k))
1070− | i >= len || cnt >= n = i
1071− | otherwise = loop (i+d) (cnt+1)
1072− where d = iter_ t i
1073−{-# INLINE splitAt #-}
1075−-- | /O(n)/ 'spanBy', applied to a predicate @p@ and text @t@, returns
1254+-- | /O(n)/ 'span', applied to a predicate @p@ and text @t@, returns
1076-- a pair whose first element is the longest prefix (possibly empty)
1255-- a pair whose first element is the longest prefix (possibly empty)
1077-- of @t@ of elements that satisfy @p@, and whose second is the
1256-- of @t@ of elements that satisfy @p@, and whose second is the
1078-- remainder of the list.
1257-- remainder of the list.
1079−spanBy :: (Char -> Bool) -> Text -> (Text, Text)
1258+span :: (Char -> Bool) -> Text -> (Text, Text)
1080−spanBy p t@(Text arr off len) = (textP arr off k, textP arr (off+k) (len-k))
1259+span p t = case span_ p t of
1081− where k = loop 0
1260+ (# hd,tl #) -> (hd,tl)
1082− loop !i | i >= len || not (p c) = i
1083− | otherwise = loop (i+d)
1084− where Iter c d = iter t i
1085−{-# INLINE spanBy #-}
1261+{-# INLINE span #-}
1087−-- | /O(n)/ 'breakBy' is like 'spanBy', but the prefix returned is
1263+-- | /O(n)/ 'break' is like 'span', but the prefix returned is
1088-- over elements that fail the predicate @p@.
1264-- over elements that fail the predicate @p@.
1089−breakBy :: (Char -> Bool) -> Text -> (Text, Text)
1265+break :: (Char -> Bool) -> Text -> (Text, Text)
1090−breakBy p = spanBy (not . p)
1266+break p = span (not . p)
1091−{-# INLINE breakBy #-}
1267+{-# INLINE break #-}
1093-- | /O(n)/ Group characters in a string according to a predicate.
1269-- | /O(n)/ Group characters in a string according to a predicate.
1094groupBy :: (Char -> Char -> Bool) -> Text -> [Text]
1270groupBy :: (Char -> Char -> Bool) -> Text -> [Text]
@@ -1108,7 +1284,7 @@ findAIndexOrEnd q t@(Text _arr _off len) = go 0
1108 where go !i | i >= len || q c = i
1284 where go !i | i >= len || q c = i
1109 | otherwise = go (i+d)
1285 | otherwise = go (i+d)
1110 where Iter c d = iter t i
1286 where Iter c d = iter t i
1112-- | /O(n)/ Group characters in a string by equality.
1288-- | /O(n)/ Group characters in a string by equality.
1113group :: Text -> [Text]
1289group :: Text -> [Text]
1114group = groupBy (==)
1290group = groupBy (==)
@@ -1132,36 +1308,43 @@ tails t | null t = [empty]
1132-- copies to create substrings; they just construct new 'Text's that
1308-- copies to create substrings; they just construct new 'Text's that
1133-- are slices of the original.
1309-- are slices of the original.
1135−-- | /O(m+n)/ Break a 'Text' into pieces separated by the first
1311+-- | /O(m+n)/ Break a 'Text' into pieces separated by the first 'Text'
1136−-- 'Text' argument, consuming the delimiter. An empty delimiter is
1312+-- argument (which cannot be empty), consuming the delimiter. An empty
1137−-- invalid, and will cause an error to be raised.
1313+-- delimiter is invalid, and will cause an error to be raised.
1139-- Examples:
1315-- Examples:
1141−-- > split "\r\n" "a\r\nb\r\nd\r\ne" == ["a","b","d","e"]
1317+-- > splitOn "\r\n" "a\r\nb\r\nd\r\ne" == ["a","b","d","e"]
1142−-- > split "aaa" "aaaXaaaXaaaXaaa" == ["","X","X","X",""]
1318+-- > splitOn "aaa" "aaaXaaaXaaaXaaa" == ["","X","X","X",""]
1143−-- > split "x" "x" == ["",""]
1319+-- > splitOn "x" "x" == ["",""]
1147−-- > intercalate s . split s == id
1323+-- > intercalate s . splitOn s == id
1148−-- > split (singleton c) == splitBy (==c)
1324+-- > splitOn (singleton c) == split (==c)
1326+-- (Note: the string @s@ to split on above cannot be empty.)
1150-- In (unlikely) bad cases, this function's time complexity degrades
1328-- In (unlikely) bad cases, this function's time complexity degrades
1151-- towards /O(n*m)/.
1329-- towards /O(n*m)/.
1152−split :: Text -> Text -> [Text]
1330+splitOn :: Text
1331+ -- ^ String to split on. If this string is empty, an error
1153−split pat@(Text _ _ l) src@(Text arr off len)
1336+splitOn pat@(Text _ _ l) src@(Text arr off len)
1154− | l <= 0 = emptyError "split"
1337+ | l <= 0 = emptyError "splitOn"
1155− | isSingleton pat = splitBy (== unsafeHead pat) src
1338+ | isSingleton pat = split (== unsafeHead pat) src
1156 | otherwise = go 0 (indices pat src)
1339 | otherwise = go 0 (indices pat src)
1158− go !s (x:xs) = textP arr (s+off) (x-s) : go (x+l) xs
1341+ go !s (x:xs) = text arr (s+off) (x-s) : go (x+l) xs
1159− go s _ = [textP arr (s+off) (len-s)]
1342+ go s _ = [text arr (s+off) (len-s)]
1160−{-# INLINE [1] split #-}
1343+{-# INLINE [1] splitOn #-}
1162{-# RULES
1345{-# RULES
1163−"TEXT split/singleton -> splitBy/==" [~1] forall c t.
1346+"TEXT splitOn/singleton -> split/==" [~1] forall c t.
1164− split (singleton c) t = splitBy (==c) t
1347+ splitOn (singleton c) t = split (==c) t
1167-- | /O(n)/ Splits a 'Text' into components delimited by separators,
1350-- | /O(n)/ Splits a 'Text' into components delimited by separators,
@@ -1169,15 +1352,15 @@ split pat@(Text _ _ l) src@(Text arr off len)
1169-- resulting components do not contain the separators. Two adjacent
1352-- resulting components do not contain the separators. Two adjacent
1170-- separators result in an empty component in the output. eg.
1353-- separators result in an empty component in the output. eg.
1172−-- > splitBy (=='a') "aabbaca" == ["","","bb","c",""]
1355+-- > split (=='a') "aabbaca" == ["","","bb","c",""]
1173−-- > splitBy (=='a') "" == [""]
1356+-- > split (=='a') "" == [""]
1174−splitBy :: (Char -> Bool) -> Text -> [Text]
1357+split :: (Char -> Bool) -> Text -> [Text]
1175−splitBy _ t@(Text _off _arr 0) = [t]
1358+split _ t@(Text _off _arr 0) = [t]
1176−splitBy p t = loop t
1359+split p t = loop t
1177 where loop s | null s' = [l]
1360 where loop s | null s' = [l]
1178 | otherwise = l : loop (unsafeTail s')
1361 | otherwise = l : loop (unsafeTail s')
1179− where (l, s') = breakBy p s
1362+ where (# l, s' #) = span_ (not . p) s
1180−{-# INLINE splitBy #-}
1363+{-# INLINE split #-}
1182-- | /O(n)/ Splits a 'Text' into components of length @k@. The last
1365-- | /O(n)/ Splits a 'Text' into components of length @k@. The last
1183-- element may be shorter than the other chunks, depending on the
1366-- element may be shorter than the other chunks, depending on the
@@ -1199,21 +1382,21 @@ chunksOf k = go
1199-------------------------------------------------------------------------------
1382-------------------------------------------------------------------------------
1200-- ** Searching with a predicate
1383-- ** Searching with a predicate
1202−-- | /O(n)/ The 'findBy' function takes a predicate and a 'Text', and
1385+-- | /O(n)/ The 'find' function takes a predicate and a 'Text', and
1203−-- returns the first element in matching the predicate, or 'Nothing'
1386+-- returns the first element matching the predicate, or 'Nothing' if
1204−-- if there is no such element.
1387+-- there is no such element.
1205−findBy :: (Char -> Bool) -> Text -> Maybe Char
1388+find :: (Char -> Bool) -> Text -> Maybe Char
1206−findBy p t = S.findBy p (stream t)
1389+find p t = S.findBy p (stream t)
1207−{-# INLINE findBy #-}
1390+{-# INLINE find #-}
1209−-- | /O(n)/ The 'partitionBy' function takes a predicate and a 'Text',
1392+-- | /O(n)/ The 'partition' function takes a predicate and a 'Text',
1210-- and returns the pair of 'Text's with elements which do and do not
1393-- and returns the pair of 'Text's with elements which do and do not
1211-- satisfy the predicate, respectively; i.e.
1394-- satisfy the predicate, respectively; i.e.
1213−-- > partitionBy p t == (filter p t, filter (not . p) t)
1396+-- > partition p t == (filter p t, filter (not . p) t)
1214−partitionBy :: (Char -> Bool) -> Text -> (Text, Text)
1397+partition :: (Char -> Bool) -> Text -> (Text, Text)
1215−partitionBy p t = (filter p t, filter (not . p) t)
1398+partition p t = (filter p t, filter (not . p) t)
1216−{-# INLINE partitionBy #-}
1399+{-# INLINE partition #-}
1218-- | /O(n)/ 'filter', applied to a predicate and a 'Text',
1401-- | /O(n)/ 'filter', applied to a predicate and a 'Text',
1219-- returns a 'Text' containing those characters that satisfy the
1402-- returns a 'Text' containing those characters that satisfy the
@@ -1229,39 +1412,40 @@ filter p t = unstream (S.filter p (stream t))
1230-- Examples:
1413-- Examples:
1232−-- > break "::" "a::b::c" ==> ("a", "::b::c")
1415+-- > breakOn "::" "a::b::c" ==> ("a", "::b::c")
1233−-- > break "/" "foobar" ==> ("foobar", "")
1416+-- > breakOn "/" "foobar" ==> ("foobar", "")
1237-- > append prefix match == haystack
1420-- > append prefix match == haystack
1238−-- > where (prefix, match) = break needle haystack
1421+-- > where (prefix, match) = breakOn needle haystack
1240-- If you need to break a string by a substring repeatedly (e.g. you
1423-- If you need to break a string by a substring repeatedly (e.g. you
1241−-- want to break on every instance of a substring), use 'find'
1424+-- want to break on every instance of a substring), use 'breakOnAll'
1242-- instead, as it has lower startup overhead.
1425-- instead, as it has lower startup overhead.
1244-- In (unlikely) bad cases, this function's time complexity degrades
1427-- In (unlikely) bad cases, this function's time complexity degrades
1245-- towards /O(n*m)/.
1428-- towards /O(n*m)/.
1246−break :: Text -> Text -> (Text, Text)
1429+breakOn :: Text -> Text -> (Text, Text)
1247−break pat src@(Text arr off len)
1430+breakOn pat src@(Text arr off len)
1248− | null pat = emptyError "break"
1431+ | null pat = emptyError "breakOn"
1249 | otherwise = case indices pat src of
1432 | otherwise = case indices pat src of
1250 [] -> (src, empty)
1433 [] -> (src, empty)
1251− (x:_) -> (textP arr off x, textP arr (off+x) (len-x))
1434+ (x:_) -> (text arr off x, text arr (off+x) (len-x))
1252−{-# INLINE break #-}
1435+{-# INLINE breakOn #-}
1254−-- | /O(n+m)/ Similar to 'break', but searches from the end of the string.
1437+-- | /O(n+m)/ Similar to 'breakOn', but searches from the end of the
1256-- The first element of the returned tuple is the prefix of @haystack@
1440-- The first element of the returned tuple is the prefix of @haystack@
1257-- up to and including the last match of @needle@. The second is the
1441-- up to and including the last match of @needle@. The second is the
1258-- remainder of @haystack@, following the match.
1442-- remainder of @haystack@, following the match.
1260−-- > breakEnd "::" "a::b::c" ==> ("a::b::", "c")
1444+-- > breakOnEnd "::" "a::b::c" ==> ("a::b::", "c")
1261−breakEnd :: Text -> Text -> (Text, Text)
1445+breakOnEnd :: Text -> Text -> (Text, Text)
1446+breakOnEnd pat src = (reverse b, reverse a)
1262−breakEnd pat src = let (a,b) = break (reverse pat) (reverse src)
1447+ where (a,b) = breakOn (reverse pat) (reverse src)
1263− in (reverse b, reverse a)
1264−{-# INLINE breakEnd #-}
1448+{-# INLINE breakOnEnd #-}
1266-- | /O(n+m)/ Find all non-overlapping instances of @needle@ in
1450-- | /O(n+m)/ Find all non-overlapping instances of @needle@ in
1267-- @haystack@. Each element of the returned list consists of a pair:
1451-- @haystack@. Each element of the returned list consists of a pair:
@@ -1272,25 +1456,25 @@ breakEnd pat src = let (a,b) = break (reverse pat) (reverse src)
1273-- Examples:
1457-- Examples:
1275−-- > find "::" ""
1459+-- > breakOnAll "::" ""
1276-- > ==> []
1460-- > ==> []
1277−-- > find "/" "a/b/c/"
1461+-- > breakOnAll "/" "a/b/c/"
1278-- > ==> [("a", "/b/c/"), ("a/b", "/c/"), ("a/b/c", "/")]
1462-- > ==> [("a", "/b/c/"), ("a/b", "/c/"), ("a/b/c", "/")]
1280-- In (unlikely) bad cases, this function's time complexity degrades
1464-- In (unlikely) bad cases, this function's time complexity degrades
1281-- towards /O(n*m)/.
1465-- towards /O(n*m)/.
1283-- The @needle@ parameter may not be empty.
1467-- The @needle@ parameter may not be empty.
1284−find :: Text -- ^ @needle@ to search for
1468+breakOnAll :: Text -- ^ @needle@ to search for
1285− -> Text -- ^ @haystack@ in which to search
1469+ -> Text -- ^ @haystack@ in which to search
1286− -> [(Text, Text)]
1470+ -> [(Text, Text)]
1287−find pat src@(Text arr off slen)
1471+breakOnAll pat src@(Text arr off slen)
1288− | null pat = emptyError "find"
1472+ | null pat = emptyError "breakOnAll"
1289 | otherwise = L.map step (indices pat src)
1473 | otherwise = L.map step (indices pat src)
1291 step x = (chunk 0 x, chunk x (slen-x))
1475 step x = (chunk 0 x, chunk x (slen-x))
1292− chunk !n !l = textP arr (n+off) l
1476+ chunk !n !l = text arr (n+off) l
1293−{-# INLINE find #-}
1477+{-# INLINE breakOnAll #-}
1295-------------------------------------------------------------------------------
1479-------------------------------------------------------------------------------
1296-- ** Indexing 'Text's
1480-- ** Indexing 'Text's
@@ -1309,7 +1493,7 @@ find pat src@(Text arr off slen)
1309-- For example, suppose you have a string that you want to split on
1493-- For example, suppose you have a string that you want to split on
1310-- the substring @\"::\"@, such as @\"foo::bar::quux\"@. Instead of
1494-- the substring @\"::\"@, such as @\"foo::bar::quux\"@. Instead of
1311-- searching for the index of @\"::\"@ and taking the substrings
1495-- searching for the index of @\"::\"@ and taking the substrings
1312−-- before and after that index, you would instead use @find \"::\"@.
1496+-- before and after that index, you would instead use @breakOnAll \"::\"@.
1314-- | /O(n)/ 'Text' index (subscript) operator, starting from 0.
1498-- | /O(n)/ 'Text' index (subscript) operator, starting from 0.
1315index :: Text -> Int -> Char
1499index :: Text -> Int -> Char
@@ -1356,7 +1540,7 @@ countChar c t = S.countChar c (stream t)
1356-- equivalent to a pair of 'unpack' operations.
1540-- equivalent to a pair of 'unpack' operations.
1357zip :: Text -> Text -> [(Char,Char)]
1541zip :: Text -> Text -> [(Char,Char)]
1358zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1542zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1359−{-# INLINE [0] zip #-}
1543+{-# INLINE zip #-}
1361-- | /O(n)/ 'zipWith' generalises 'zip' by zipping with the function
1545-- | /O(n)/ 'zipWith' generalises 'zip' by zipping with the function
1362-- given as the first argument, instead of a tupling function.
1546-- given as the first argument, instead of a tupling function.
@@ -1364,7 +1548,7 @@ zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1364zipWith :: (Char -> Char -> Char) -> Text -> Text -> Text
1548zipWith :: (Char -> Char -> Char) -> Text -> Text -> Text
1365zipWith f t1 t2 = unstream (S.zipWith g (stream t1) (stream t2))
1549zipWith f t1 t2 = unstream (S.zipWith g (stream t1) (stream t2))
1366 where g a b = safe (f a b)
1550 where g a b = safe (f a b)
1367−{-# INLINE [0] zipWith #-}
1551+{-# INLINE zipWith #-}
1369-- | /O(n)/ Breaks a 'Text' up into a list of words, delimited by 'Char's
1553-- | /O(n)/ Breaks a 'Text' up into a list of words, delimited by 'Char's
1370-- representing white space.
1554-- representing white space.
@@ -1390,7 +1574,7 @@ lines ps | null ps = []
1390 | otherwise = h : if null t
1574 | otherwise = h : if null t
1392 else lines (unsafeTail t)
1576 else lines (unsafeTail t)
1393− where (h,t) = spanBy (/= '\n') ps
1577+ where (# h,t #) = span_ (/= '\n') ps
1394{-# INLINE lines #-}
1578{-# INLINE lines #-}
@@ -1436,8 +1620,6 @@ isPrefixOf a@(Text _ _ alen) b@(Text _ _ blen) =
1436{-# RULES
1620{-# RULES
1437"TEXT isPrefixOf -> fused" [~1] forall s t.
1621"TEXT isPrefixOf -> fused" [~1] forall s t.
1438 isPrefixOf s t = S.isPrefixOf (stream s) (stream t)
1622 isPrefixOf s t = S.isPrefixOf (stream s) (stream t)
1439−"TEXT isPrefixOf -> unfused" [1] forall s t.
1440− S.isPrefixOf (stream s) (stream t) = isPrefixOf s t
1443-- | /O(n)/ The 'isSuffixOf' function takes two 'Text's and returns
1625-- | /O(n)/ The 'isSuffixOf' function takes two 'Text's and returns
@@ -1471,12 +1653,13 @@ isInfixOf needle haystack
1471-------------------------------------------------------------------------------
1653-------------------------------------------------------------------------------
1472-- * View patterns
1654-- * View patterns
1474−-- | /O(n)/ Returns the suffix of the second string if its prefix
1656+-- | /O(n)/ Return the suffix of the second string if its prefix
1475−-- matches the first.
1657+-- matches the entire first string.
1477-- Examples:
1659-- Examples:
1479-- > stripPrefix "foo" "foobar" == Just "bar"
1661-- > stripPrefix "foo" "foobar" == Just "bar"
1662+-- > stripPrefix "" "baz" == Just "baz"
1480-- > stripPrefix "foo" "quux" == Nothing
1663-- > stripPrefix "foo" "quux" == Nothing
1482-- This is particularly useful with the @ViewPatterns@ extension to
1665-- This is particularly useful with the @ViewPatterns@ extension to
@@ -1487,18 +1670,42 @@ isInfixOf needle haystack
1488-- > fnordLength :: Text -> Int
1671-- > fnordLength :: Text -> Int
1489-- > fnordLength (stripPrefix "fnord" -> Just suf) = T.length suf
1672-- > fnordLength (stripPrefix "fnord" -> Just suf) = T.length suf
1490−-- > fnordLength _ = -1
1673+-- > fnordLength _ = -1
1491stripPrefix :: Text -> Text -> Maybe Text
1674stripPrefix :: Text -> Text -> Maybe Text
1492stripPrefix p@(Text _arr _off plen) t@(Text arr off len)
1675stripPrefix p@(Text _arr _off plen) t@(Text arr off len)
1493− | p `isPrefixOf` t = Just $! textP arr (off+plen) (len-plen)
1676+ | p `isPrefixOf` t = Just $! text arr (off+plen) (len-plen)
1494 | otherwise = Nothing
1677 | otherwise = Nothing
1679+-- | /O(n)/ Find the longest non-empty common prefix of two strings
1680+-- and return it, along with the suffixes of each string at which they
1683+-- If the strings do not have a common prefix or either one is empty,
1684+-- this function returns 'Nothing'.
1688+-- > commonPrefixes "foobar" "fooquux" == Just ("foo","bar","quux")
1689+-- > commonPrefixes "veeble" "fetzer" == Nothing
1690+-- > commonPrefixes "" "baz" == Nothing
1691+commonPrefixes :: Text -> Text -> Maybe (Text,Text,Text)
1692+commonPrefixes t0@(Text arr0 off0 len0) t1@(Text arr1 off1 len1) = go 0 0
1694+ go !i !j | i < len0 && j < len1 && a == b = go (i+d0) (j+d1)
1695+ | i > 0 = Just (Text arr0 off0 i,
1696+ text arr0 (off0+i) (len0-i),
1697+ text arr1 (off1+j) (len1-j))
1698+ | otherwise = Nothing
1699+ where Iter a d0 = iter t0 i
1700+ Iter b d1 = iter t1 j
1496−-- | /O(n)/ Returns the prefix of the second string if its suffix
1702+-- | /O(n)/ Return the prefix of the second string if its suffix
1497−-- matches the first.
1703+-- matches the entire first string.
1499-- Examples:
1705-- Examples:
1501-- > stripSuffix "bar" "foobar" == Just "foo"
1707-- > stripSuffix "bar" "foobar" == Just "foo"
1708+-- > stripSuffix "" "baz" == Just "baz"
1502-- > stripSuffix "foo" "quux" == Nothing
1709-- > stripSuffix "foo" "quux" == Nothing
1504-- This is particularly useful with the @ViewPatterns@ extension to
1711-- This is particularly useful with the @ViewPatterns@ extension to
@@ -1509,10 +1716,10 @@ stripPrefix p@(Text _arr _off plen) t@(Text arr off len)
1510-- > quuxLength :: Text -> Int
1717-- > quuxLength :: Text -> Int
1511-- > quuxLength (stripSuffix "quux" -> Just pre) = T.length pre
1718-- > quuxLength (stripSuffix "quux" -> Just pre) = T.length pre
1512−-- > quuxLength _ = -1
1719+-- > quuxLength _ = -1
1513stripSuffix :: Text -> Text -> Maybe Text
1720stripSuffix :: Text -> Text -> Maybe Text
1514stripSuffix p@(Text _arr _off plen) t@(Text arr off len)
1721stripSuffix p@(Text _arr _off plen) t@(Text arr off len)
1515− | p `isSuffixOf` t = Just $! textP arr off (len-plen)
1722+ | p `isSuffixOf` t = Just $! text arr off (len-plen)
1516 | otherwise = Nothing
1723 | otherwise = Nothing
1518-- | Add a list of non-negative numbers. Errors out on overflow.
1725-- | Add a list of non-negative numbers. Errors out on overflow.
@@ -1529,3 +1736,20 @@ emptyError fun = P.error $ "Data.Text." ++ fun ++ ": empty input"
1530overflowError :: String -> a
1737overflowError :: String -> a
1531overflowError fun = P.error $ "Data.Text." ++ fun ++ ": size overflow"
1738overflowError fun = P.error $ "Data.Text." ++ fun ++ ": size overflow"
1740+-- | /O(n)/ Make a distinct copy of the given string, sharing no
1741+-- storage with the original string.
1743+-- As an example, suppose you read a large string, of which you need
1744+-- only a small portion. If you do not use 'copy', the entire original
1745+-- array will be kept alive in memory by the smaller string. Making a
1746+-- copy \"breaks the link\" to the original array, allowing it to be
1747+-- garbage collected if there are no other live references to it.
1748+copy :: Text -> Text
1749+copy (Text arr off len) = Text (A.run go) 0 len
1751+ go :: ST s (A.MArray s)
1754+ A.copyI marr 0 arr off len