@@ -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
@@ -167,92 +176,110 @@ module Data.Text
167 , partition
176 , partition
169 -- , findSubstring
178 -- , findSubstring
171 -- * Indexing
180 -- * Indexing
172 -- $index
181 -- $index
174 , findIndex
183 , findIndex
177− -- * Zipping and unzipping
186+ -- * Zipping
179 , zipWith
188 , zipWith
181 -- -* Ordered text
190 -- -* Ordered text
182 -- , sort
191 -- , sort
193+ -- * Low level operations
185import Prelude (Char, Bool(..), Int, Maybe(..), String,
198import Prelude (Char, Bool(..), Int, Maybe(..), String,
186 Eq(..), Ord(..), Ordering(..), (++),
199 Eq(..), Ord(..), Ordering(..), (++),
187− Read(..), Show(..),
200+ Read(..),
188− (&&), (||), (+), (-), (.), ($), ($!), (>>), (*),
201+ (&&), (||), (+), (-), (.), ($), ($!), (>>),
189− div, maxBound, not, return, otherwise)
202+ not, return, otherwise, quot)
190#if defined(HAVE_DEEPSEQ)
203#if defined(HAVE_DEEPSEQ)
191−import Control.DeepSeq (NFData)
204+import Control.DeepSeq (NFData(rnf))
193#if defined(ASSERTS)
206#if defined(ASSERTS)
194import Control.Exception (assert)
207import Control.Exception (assert)
196import Data.Char (isSpace)
209import Data.Char (isSpace)
197−import Data.Data (Data(gfoldl, toConstr, gunfold, dataTypeOf))
210+import Data.Data (Data(gfoldl, toConstr, gunfold, dataTypeOf), constrIndex,
198−#if __GLASGOW_HASKELL__ >= 612
211+ Constr, mkConstr, DataType, mkDataType, Fixity(Prefix))
199−import Data.Data (mkNoRepType)
201−import Data.Data (mkNorepType)
203import Control.Monad (foldM)
212import Control.Monad (foldM)
213+import Control.Monad.ST (ST)
204import qualified Data.Text.Array as A
214import qualified Data.Text.Array as A
205import qualified Data.List as L
215import qualified Data.List as L
216+import Data.Binary (Binary(get, put))
206import Data.Monoid (Monoid(..))
217import Data.Monoid (Monoid(..))
218+#if MIN_VERSION_base(4,9,0)
219+import Data.Semigroup (Semigroup(..))
207import Data.String (IsString(..))
221import Data.String (IsString(..))
208−import qualified Data.Text.Fusion as S
222+import qualified Data.Text.Internal.Fusion as S
209−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)
210−import Data.Text.Fusion (stream, reverseStream, unstream)
225+import Data.Text.Internal.Fusion (stream, reverseStream, unstream)
211−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#)
212import qualified Prelude as P
229import qualified Prelude as P
213import Data.Text.Unsafe (Iter(..), iter, iter_, lengthWord16, reverseIter,
230import Data.Text.Unsafe (Iter(..), iter, iter_, lengthWord16, reverseIter,
214− unsafeHead, unsafeTail)
231+ reverseIter_, unsafeHead, unsafeTail)
215−import Data.Text.UnsafeChar (unsafeChr)
232+import Data.Text.Internal.Unsafe.Char (unsafeChr)
216−import qualified Data.Text.Util as U
233+import qualified Data.Text.Internal.Functions as F
217−import qualified Data.Text.Encoding.Utf16 as U16
234+import qualified Data.Text.Internal.Encoding.Utf16 as U16
218−import Data.Text.Search (indices)
235+import Data.Text.Internal.Search (indices)
219#if defined(__HADDOCK__)
236#if defined(__HADDOCK__)
220import Data.ByteString (ByteString)
237import Data.ByteString (ByteString)
221import qualified Data.Text.Lazy as L
238import qualified Data.Text.Lazy as L
222import 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)
225-- $strict
249-- $strict
227-- This package provides both strict and lazy 'Text' types. The
251-- This package provides both strict and lazy 'Text' types. The
228−-- 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
229−-- type is provided by the 'Data.Text.Lazy' package. Internally, the
253+-- type is provided by the "Data.Text.Lazy" module. Internally, the
230-- lazy @Text@ type consists of a list of strict chunks.
254-- lazy @Text@ type consists of a list of strict chunks.
232-- The strict 'Text' type requires that an entire string fit into
256-- The strict 'Text' type requires that an entire string fit into
233−-- 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
234−-- strings that are larger than memory using a small memory footprint.
258+-- streaming strings that are larger than memory using a small memory
235−-- In many cases, the overhead of chunked streaming makes the lazy
259+-- footprint. In many cases, the overhead of chunked streaming makes
236−-- @Text@ type slower than its strict counterpart, but this is not
260+-- the lazy 'Data.Text.Lazy.Text' type slower than its strict
237−-- always the case. Sometimes, the time complexity of a function in
261+-- counterpart, but this is not always the case. Sometimes, the time
238−-- 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
239−-- internal structures.
263+-- other, due to their differing internal structures.
241-- Each module provides an almost identical API, with the main
265-- Each module provides an almost identical API, with the main
242-- difference being that the strict module uses 'Int' values for
266-- difference being that the strict module uses 'Int' values for
243−-- lengths and counts, while the lazy module uses 'Int64' lengths.
267+-- lengths and counts, while the lazy module uses 'Data.Int.Int64'
245-- $replacement
270-- $replacement
247-- 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
248−-- 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 >.
249−-- <http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=35>. As
250−-- 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
251−-- 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 >)
252−-- (§3.4, definition D10) as 'Char' values, including code points
279+-- as 'Char' values, including code points from this invalid range.
253−-- from this invalid range. This means that there are some 'Char'
280+-- This means that there are some 'Char' values that are not valid
254−-- values that are not valid Unicode scalar values, and the functions
281+-- Unicode scalar values, and the functions in this module must handle
255−-- in this module must handle those cases.
282+-- those cases.
257-- Within this module, many functions construct a 'Text' from one or
284-- Within this module, many functions construct a 'Text' from one or
258-- more 'Char' values. Those functions will substitute 'Char' values
285-- more 'Char' values. Those functions will substitute 'Char' values
@@ -266,8 +293,8 @@ import Data.Int (Int64)
266-- 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
267-- code points, and so cannot be represented. The functions replace
294-- code points, and so cannot be represented. The functions replace
268-- invalid scalar values, instead of dropping them, as a security
295-- invalid scalar values, instead of dropping them, as a security
269−-- measure. For details, see Unicode Technical Report 36, §3.5:
296+-- measure. For details, see
270−-- <http://unicode.org/reports/tr36/#Deletion_of_Noncharacters>)
297+-- <http://unicode.org/reports/tr36/#Deletion_of_Noncharacters Unicode Technical Report 36, §3.5 >.)
272-- $fusion
299-- $fusion
@@ -285,11 +312,11 @@ import Data.Int (Int64)
285-- > countChars = T.length . T.toUpper . E.decodeUtf8
312-- > countChars = T.length . T.toUpper . E.decodeUtf8
287-- From the type signatures involved, this looks like it should
314-- From the type signatures involved, this looks like it should
288−-- allocate one 'ByteString' value, and two 'Text' values. However,
315+-- allocate one 'Data.ByteString.ByteString' value, and two 'Text'
289−-- when a module is compiled with optimisation enabled under GHC, the
316+-- values. However, when a module is compiled with optimisation
290−-- two intermediate 'Text' values will be optimised away, and the
317+-- enabled under GHC, the two intermediate 'Text' values will be
291−-- 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
292−-- 'ByteString'.
319+-- loop over the source 'Data.ByteString.ByteString'.
294-- 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
295-- phrase \"Subject to fusion\".
322-- phrase \"Subject to fusion\".
@@ -303,46 +330,83 @@ instance Eq Text where
303instance Ord Text where
330instance Ord Text where
304 compare = compareText
331 compare = compareText
306−instance Show Text where
307− showsPrec p ps r = showsPrec p (unpack ps) r
309instance Read Text where
333instance Read Text where
310 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
312instance Monoid Text where
344instance Monoid Text where
313 mempty = empty
345 mempty = empty
346+#if MIN_VERSION_base(4,9,0)
347+ mappend = (<>) -- future-proof definition
314 mappend = append
349 mappend = append
315 mconcat = concat
351 mconcat = concat
317instance IsString Text where
353instance IsString Text where
318 fromString = pack
354 fromString = pack
356+#if __GLASGOW_HASKELL__ >= 708
357+instance Exts.IsList Text where
358+ type Item Text = Char
320#if defined(HAVE_DEEPSEQ)
363#if defined(HAVE_DEEPSEQ)
321−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
324−-- This instance preserves data abstraction at the cost of inefficiency.
375+-- | This instance preserves data abstraction at the cost of inefficiency.
325-- We omit reflection services for the sake of data abstraction.
376-- We omit reflection services for the sake of data abstraction.
327−-- This instance was created by copying the behavior of Data.Set and
328−-- Data.Map. If you feel a mistake has been made, please feel free to
329−-- submit improvements.
331−-- Original discussion is archived here:
378+-- This instance was created by copying the updated behavior of
332−
379+-- @"Data.Set".@'Data.Set.Set' and @"Data.Map".@'Data.Map.Map'. If you
333−-- "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:
334−-- 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, ... >
336instance Data Text where
390instance Data Text where
337 gfoldl f z txt = z pack `f` (unpack txt)
391 gfoldl f z txt = z pack `f` (unpack txt)
338− toConstr _ = P.error "Data.Text.Text.toConstr"
392+ toConstr _ = packConstr
393+ gunfold k z c = case constrIndex c of
339− gunfold _ _ = P.error "Data.Text.Text.gunfold"
395+ _ -> P.error "gunfold"
340−#if __GLASGOW_HASKELL__ >= 612
396+ dataTypeOf _ = textDataType
341− dataTypeOf _ = mkNoRepType "Data.Text.Text"
397+
342−#else
398+#if MIN_VERSION_base(4,7,0)
343− 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]
346-- | /O(n)/ Compare two 'Text' values lexicographically.
410-- | /O(n)/ Compare two 'Text' values lexicographically.
347compareText :: Text -> Text -> Ordering
411compareText :: Text -> Text -> Ordering
348compareText ta@(Text _arrA _offA lenA) tb@(Text _arrB _offB lenB)
412compareText ta@(Text _arrA _offA lenA) tb@(Text _arrB _offB lenB)
@@ -363,20 +427,9 @@ compareText ta@(Text _arrA _offA lenA) tb@(Text _arrB _offB lenB)
363-- | /O(n)/ Convert a 'String' into a 'Text'. Subject to
427-- | /O(n)/ Convert a 'String' into a 'Text'. Subject to
364-- fusion. Performs replacement on invalid scalar values.
428-- fusion. Performs replacement on invalid scalar values.
365pack :: String -> Text
429pack :: String -> Text
366−pack = unstream . S.streamList . L.map safe
430+pack = unstream . S.map safe . S.streamList
367{-# INLINE [1] pack #-}
431{-# INLINE [1] pack #-}
369−-- | /O(n)/ Convert a Text into a String. Subject to fusion.
370−unpack :: Text -> String
371−unpack = S.unstreamList . stream
372−{-# INLINE [1] unpack #-}
374−-- | /O(1)/ Convert a character into a Text. Subject to fusion.
375−-- Performs replacement on invalid scalar values.
376−singleton :: Char -> Text
377−singleton = unstream . S.singleton . safe
378−{-# INLINE [1] singleton #-}
380-- -----------------------------------------------------------------------------
433-- -----------------------------------------------------------------------------
381-- * Basic functions
434-- * Basic functions
@@ -407,12 +460,13 @@ append a@(Text arr1 off1 len1) b@(Text arr2 off2 len2)
407 | otherwise = overflowError "append"
460 | otherwise = overflowError "append"
409 len = len1+len2
462 len = len1+len2
463+ x :: ST s (A.MArray s)
411 arr <- A.new len
465 arr <- A.new len
412 A.copyI arr 0 arr1 off1 len1
466 A.copyI arr 0 arr1 off1 len1
413 A.copyI arr len1 arr2 off2 len
467 A.copyI arr len1 arr2 off2 len
414 return arr
468 return arr
415−{-# INLINE append #-}
469+{-# NOINLINE append #-}
418"TEXT append -> fused" [~1] forall t1 t2.
472"TEXT append -> fused" [~1] forall t1 t2.
@@ -432,8 +486,8 @@ head t = S.head (stream t)
432uncons :: Text -> Maybe (Char, Text)
486uncons :: Text -> Maybe (Char, Text)
433uncons t@(Text arr off len)
487uncons t@(Text arr off len)
434 | len <= 0 = Nothing
488 | len <= 0 = Nothing
489+ | otherwise = Just $ let !(Iter c d) = iter t 0
435− | otherwise = Just (c, textP arr (off+d) (len-d))
490+ in (c, text arr (off+d) (len-d))
436− where Iter c d = iter t 0
437{-# INLINE [1] uncons #-}
491{-# INLINE [1] uncons #-}
439-- | Lifted from Control.Arrow and specialized.
493-- | Lifted from Control.Arrow and specialized.
@@ -463,7 +517,7 @@ last (Text arr off len)
463tail :: Text -> Text
517tail :: Text -> Text
464tail t@(Text arr off len)
518tail t@(Text arr off len)
465 | len <= 0 = emptyError "tail"
519 | len <= 0 = emptyError "tail"
466− | otherwise = textP arr (off+d) (len-d)
520+ | otherwise = text arr (off+d) (len-d)
467 where d = iter_ t 0
521 where d = iter_ t 0
468{-# INLINE [1] tail #-}
522{-# INLINE [1] tail #-}
@@ -478,8 +532,8 @@ tail t@(Text arr off len)
478-- be non-empty. Subject to fusion.
532-- be non-empty. Subject to fusion.
479init :: Text -> Text
533init :: Text -> Text
480init (Text arr off len) | len <= 0 = emptyError "init"
534init (Text arr off len) | len <= 0 = emptyError "init"
481− | n >= 0xDC00 && n <= 0xDFFF = textP arr off (len-2)
535+ | n >= 0xDC00 && n <= 0xDFFF = text arr off (len-2)
482− | otherwise = textP arr off (len-1)
536+ | otherwise = text arr off (len-1)
484 n = A.unsafeIndex arr (off+len-1)
538 n = A.unsafeIndex arr (off+len-1)
485{-# INLINE [1] init #-}
539{-# INLINE [1] init #-}
@@ -518,7 +572,9 @@ isSingleton = S.isSingleton . stream
518-- Subject to fusion.
572-- Subject to fusion.
519length :: Text -> Int
573length :: Text -> Int
520length t = S.length (stream t)
574length t = S.length (stream t)
521−{-# 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.
523-- | /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.
524-- Subject to fusion.
580-- Subject to fusion.
@@ -537,32 +593,32 @@ compareLength t n = S.compareLengthI (stream t) n
539"TEXT ==N/length -> compareLength/==EQ" [~1] forall t n.
595"TEXT ==N/length -> compareLength/==EQ" [~1] forall t n.
540− (==) (length t) n = compareLength t n == EQ
596+ eqInt (length t) n = compareLength t n == EQ
544"TEXT /=N/length -> compareLength//=EQ" [~1] forall t n.
600"TEXT /=N/length -> compareLength//=EQ" [~1] forall t n.
545− (/=) (length t) n = compareLength t n /= EQ
601+ neInt (length t) n = compareLength t n /= EQ
549"TEXT <N/length -> compareLength/==LT" [~1] forall t n.
605"TEXT <N/length -> compareLength/==LT" [~1] forall t n.
550− (<) (length t) n = compareLength t n == LT
606+ ltInt (length t) n = compareLength t n == LT
554"TEXT <=N/length -> compareLength//=GT" [~1] forall t n.
610"TEXT <=N/length -> compareLength//=GT" [~1] forall t n.
555− (<=) (length t) n = compareLength t n /= GT
611+ leInt (length t) n = compareLength t n /= GT
559"TEXT >N/length -> compareLength/==GT" [~1] forall t n.
615"TEXT >N/length -> compareLength/==GT" [~1] forall t n.
560− (>) (length t) n = compareLength t n == GT
616+ gtInt (length t) n = compareLength t n == GT
564"TEXT >=N/length -> compareLength//=LT" [~1] forall t n.
620"TEXT >=N/length -> compareLength//=LT" [~1] forall t n.
565− (>=) (length t) n = compareLength t n /= LT
621+ geInt (length t) n = compareLength t n /= LT
568-- -----------------------------------------------------------------------------
624-- -----------------------------------------------------------------------------
@@ -578,7 +634,7 @@ map f t = unstream (S.map (safe . f) (stream t))
578-- 'Text's and concatenates the list after interspersing the first
634-- 'Text's and concatenates the list after interspersing the first
579-- argument between each element of the list.
635-- argument between each element of the list.
580intercalate :: Text -> [Text] -> Text
636intercalate :: Text -> [Text] -> Text
581−intercalate t = concat . (U.intersperse t)
637+intercalate t = concat . (F.intersperse t)
582{-# INLINE intercalate #-}
638{-# INLINE intercalate #-}
584-- | /O(n)/ The 'intersperse' function takes a character and places it
640-- | /O(n)/ The 'intersperse' function takes a character and places it
@@ -593,16 +649,59 @@ reverse :: Text -> Text
593reverse t = S.reverse (stream t)
649reverse t = S.reverse (stream t)
594{-# INLINE reverse #-}
650{-# INLINE reverse #-}
596−-- | /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"
598-- In (unlikely) bad cases, this function's time complexity degrades
673-- In (unlikely) bad cases, this function's time complexity degrades
599-- towards /O(n*m)/.
674-- towards /O(n*m)/.
600−replace :: Text -- ^ Text to search for
675+replace :: Text
601− -> Text -- ^ Replacement text
676+ -- ^ @needle@ to search for. If this string is empty, an
677+ -- error will occur.
679+ -- ^ @replacement@ to replace @needle@ with.
602− -> Text -- ^ Input text
680+ -> Text
681+ -- ^ @haystack@ in which to search.
604−replace s d = intercalate d . splitOn s
683+replace needle@(Text _ _ neeLen)
605−{-# 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
607-- ----------------------------------------------------------------------------
706-- ----------------------------------------------------------------------------
608-- ** Case conversions (folds)
707-- ** Case conversions (folds)
@@ -621,13 +720,13 @@ replace s d = intercalate d . splitOn s
621-- /Note/: In some languages, case conversion is a locale- and
720-- /Note/: In some languages, case conversion is a locale- and
622-- context-dependent operation. The case conversion functions in this
721-- context-dependent operation. The case conversion functions in this
623-- module are /not/ locale sensitive. Programs that require locale
722-- module are /not/ locale sensitive. Programs that require locale
624−-- sensitivity should use appropriate versions of the case mapping
723+-- sensitivity should use appropriate versions of the
625−-- functions from the @text-icu@ package:
626−-- <http://hackage.haskell.org/package/text-icu>
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 >.
628−-- | /O(n)/ Convert a string to folded case. This function is mainly
726+-- | /O(n)/ Convert a string to folded case. Subject to fusion.
629−-- useful for performing caseless (also known as case insensitive)
728+-- This function is mainly useful for performing caseless (also known
630−-- string comparisons.
729+-- as case insensitive) string comparisons.
632-- 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:
@@ -642,25 +741,51 @@ replace s d = intercalate d . splitOn s
642-- instead of itself.
741-- instead of itself.
643toCaseFold :: Text -> Text
742toCaseFold :: Text -> Text
644toCaseFold t = unstream (S.toCaseFold (stream t))
743toCaseFold t = unstream (S.toCaseFold (stream t))
645−{-# INLINE [0] toCaseFold #-}
744+{-# INLINE toCaseFold #-}
647-- | /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.
648−-- conversion. The result string may be longer than the input string.
749+-- The result string may be longer than the input string. For
649−-- For instance, \"İ\" (Latin capital letter I with dot above,
750+-- instance, \"İ\" (Latin capital letter I with dot above,
650−-- 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)
651−-- by \" ̇\" (combining dot above, U+0307).
752+-- followed by \" ̇\" (combining dot above, U+0307).
652toLower :: Text -> Text
753toLower :: Text -> Text
653toLower t = unstream (S.toLower (stream t))
754toLower t = unstream (S.toLower (stream t))
654{-# INLINE toLower #-}
755{-# INLINE toLower #-}
656-- | /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.
657−-- conversion. The result string may be longer than the input string.
760+-- The result string may be longer than the input string. For
658−-- For instance, the German \"ß\" (eszett, U+00DF) maps to the
761+-- instance, the German \"ß\" (eszett, U+00DF) maps to the
659-- two-letter sequence \"SS\".
762-- two-letter sequence \"SS\".
660toUpper :: Text -> Text
763toUpper :: Text -> Text
661toUpper t = unstream (S.toUpper (stream t))
764toUpper t = unstream (S.toUpper (stream t))
662{-# 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 #-}
664-- | /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
665-- specified fill character on the right. Subject to fusion.
790-- specified fill character on the right. Subject to fusion.
666-- Performs replacement on invalid scalar values.
791-- Performs replacement on invalid scalar values.
@@ -711,7 +836,7 @@ center k c t
711 | otherwise = replicateChar l c `append` t `append` replicateChar r c
836 | otherwise = replicateChar l c `append` t `append` replicateChar r c
712 where len = length t
837 where len = length t
713 d = k - len
838 d = k - len
714− r = d `div` 2
839+ r = d `quot` 2
715 l = d - r
840 l = d - r
716{-# INLINE center #-}
841{-# INLINE center #-}
@@ -776,6 +901,7 @@ concat ts = case ts' of
777 ts' = L.filter (not . null) ts
902 ts' = L.filter (not . null) ts
778 len = sumP "concat" $ L.map lengthWord16 ts'
903 len = sumP "concat" $ L.map lengthWord16 ts'
904+ go :: ST s (A.MArray s)
780 arr <- A.new len
906 arr <- A.new len
781 let step i (Text a o l) =
907 let step i (Text a o l) =
@@ -789,13 +915,13 @@ concatMap f = concat . foldr ((:) . f) []
789{-# INLINE concatMap #-}
915{-# INLINE concatMap #-}
791-- | /O(n)/ 'any' @p@ @t@ determines whether any character in the
917-- | /O(n)/ 'any' @p@ @t@ determines whether any character in the
792−-- 'Text' @t@ satisifes the predicate @p@. Subject to fusion.
918+-- 'Text' @t@ satisfies the predicate @p@. Subject to fusion.
793any :: (Char -> Bool) -> Text -> Bool
919any :: (Char -> Bool) -> Text -> Bool
794any p t = S.any p (stream t)
920any p t = S.any p (stream t)
795{-# INLINE any #-}
921{-# INLINE any #-}
797-- | /O(n)/ 'all' @p@ @t@ determines whether all characters in the
923-- | /O(n)/ 'all' @p@ @t@ determines whether all characters in the
798−-- 'Text' @t@ satisify the predicate @p@. Subject to fusion.
924+-- 'Text' @t@ satisfy the predicate @p@. Subject to fusion.
799all :: (Char -> Bool) -> Text -> Bool
925all :: (Char -> Bool) -> Text -> Bool
800all p t = S.all p (stream t)
926all p t = S.all p (stream t)
801{-# INLINE all #-}
927{-# INLINE all #-}
@@ -883,13 +1009,13 @@ mapAccumR f z0 = second reverse . S.mapAccumL g z0 . reverseStream
883-- @t@ repeated @n@ times.
1009-- @t@ repeated @n@ times.
884replicate :: Int -> Text -> Text
1010replicate :: Int -> Text -> Text
885replicate n t@(Text a o l)
1011replicate n t@(Text a o l)
886− | n <= 0 || l <= 0 = empty
1012+ | n <= 0 || l <= 0 = empty
887− | n == 1 = t
1013+ | n == 1 = t
888− | isSingleton t = replicateChar n (unsafeHead t)
1014+ | isSingleton t = replicateChar n (unsafeHead t)
889− | n <= maxBound `div` l = Text (A.run x) 0 len
1015+ | otherwise = Text (A.run x) 0 len
890− | otherwise = overflowError "replicate"
892− len = l * n
1017+ len = l `mul` n
1018+ x :: ST s (A.MArray s)
894 arr <- A.new len
1020 arr <- A.new len
895 let loop !d !i | i >= n = return arr
1021 let loop !d !i | i >= n = return arr
@@ -940,14 +1066,16 @@ take :: Int -> Text -> Text
940take n t@(Text arr off len)
1066take n t@(Text arr off len)
941 | n <= 0 = empty
1067 | n <= 0 = empty
942 | n >= len = t
1068 | n >= len = t
943− | otherwise = Text arr off (loop 0 0)
1069+ | otherwise = text arr off (iterN n t)
946− | i >= len || cnt >= n = i
947− | otherwise = loop (i+d) (cnt+1)
949{-# 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
951{-# RULES
1079{-# RULES
952"TEXT take -> fused" [~1] forall n t.
1080"TEXT take -> fused" [~1] forall n t.
953 take n t = unstream (S.take n (stream t))
1081 take n t = unstream (S.take n (stream t))
@@ -955,6 +1083,27 @@ take n t@(Text arr off len)
955 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
958-- | /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
959-- '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@
960-- is greater than the length of the 'Text'. Subject to fusion.
1109-- is greater than the length of the 'Text'. Subject to fusion.
@@ -962,11 +1111,8 @@ drop :: Int -> Text -> Text
962drop n t@(Text arr off len)
1111drop n t@(Text arr off len)
963 | n <= 0 = t
1112 | n <= 0 = t
964 | n >= len = empty
1113 | n >= len = empty
965− | otherwise = loop 0 0
1114+ | otherwise = text arr (off+i) (len-i)
967− | i >= len || cnt >= n = Text arr (off+i) (len-i)
968− | otherwise = loop (i+d) (cnt+1)
969− where d = iter_ t i
1115+ where i = iterN n t
970{-# INLINE [1] drop #-}
1116{-# INLINE [1] drop #-}
972{-# RULES
1118{-# RULES
@@ -976,6 +1122,18 @@ drop n t@(Text arr off len)
976 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)
979-- | /O(n)/ 'takeWhile', applied to a predicate @p@ and a 'Text',
1137-- | /O(n)/ 'takeWhile', applied to a predicate @p@ and a 'Text',
980-- returns the longest prefix (possibly empty) of elements that
1138-- returns the longest prefix (possibly empty) of elements that
981-- satisfy @p@. Subject to fusion.
1139-- satisfy @p@. Subject to fusion.
@@ -983,7 +1141,7 @@ takeWhile :: (Char -> Bool) -> Text -> Text
983takeWhile p t@(Text arr off len) = loop 0
1141takeWhile p t@(Text arr off len) = loop 0
984 where loop !i | i >= len = t
1142 where loop !i | i >= len = t
985 | p c = loop (i+d)
1143 | p c = loop (i+d)
986− | otherwise = textP arr off i
1144+ | otherwise = text arr off i
987 where Iter c d = iter t i
1145 where Iter c d = iter t i
988{-# INLINE [1] takeWhile #-}
1146{-# INLINE [1] takeWhile #-}
@@ -994,6 +1152,27 @@ takeWhile p t@(Text arr off len) = loop 0
994 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
997-- | /O(n)/ 'dropWhile' @p@ @t@ returns the suffix remaining after
1176-- | /O(n)/ 'dropWhile' @p@ @t@ returns the suffix remaining after
998-- 'takeWhile' @p@ @t@. Subject to fusion.
1177-- 'takeWhile' @p@ @t@. Subject to fusion.
999dropWhile :: (Char -> Bool) -> Text -> Text
1178dropWhile :: (Char -> Bool) -> Text -> Text
@@ -1012,8 +1191,9 @@ dropWhile p t@(Text arr off len) = loop 0 0
1014-- | /O(n)/ 'dropWhileEnd' @p@ @t@ returns the prefix remaining after
1193-- | /O(n)/ 'dropWhileEnd' @p@ @t@ returns the prefix remaining after
1015−-- dropping characters that fail the predicate @p@ from the end of
1194+-- dropping characters that satisfy the predicate @p@ from the end of
1016-- @t@. Subject to fusion.
1195-- @t@. Subject to fusion.
1017-- Examples:
1197-- Examples:
1019-- > dropWhileEnd (=='.') "foo..." == "foo"
1199-- > dropWhileEnd (=='.') "foo..." == "foo"
@@ -1033,7 +1213,7 @@ dropWhileEnd p t@(Text arr off len) = loop (len-1) len
1035-- | /O(n)/ 'dropAround' @p@ @t@ returns the substring remaining after
1215-- | /O(n)/ 'dropAround' @p@ @t@ returns the substring remaining after
1036−-- dropping characters that fail the predicate @p@ from both the
1216+-- dropping characters that satisfy the predicate @p@ from both the
1037-- beginning and end of @t@. Subject to fusion.
1217-- beginning and end of @t@. Subject to fusion.
1038dropAround :: (Char -> Bool) -> Text -> Text
1218dropAround :: (Char -> Bool) -> Text -> Text
1039dropAround p = dropWhile p . dropWhileEnd p
1219dropAround p = dropWhile p . dropWhileEnd p
@@ -1068,24 +1248,16 @@ splitAt :: Int -> Text -> (Text, Text)
1068splitAt n t@(Text arr off len)
1248splitAt n t@(Text arr off len)
1069 | n <= 0 = (empty, t)
1249 | n <= 0 = (empty, t)
1070 | n >= len = (t, empty)
1250 | n >= len = (t, empty)
1071− | otherwise = (Text arr off k, Text arr (off+k) (len-k))
1251+ | otherwise = let k = iterN n t
1072− where k = loop 0 0
1252+ in (text arr off k, text arr (off+k) (len-k))
1074− | i >= len || cnt >= n = i
1075− | otherwise = loop (i+d) (cnt+1)
1076− where d = iter_ t i
1077−{-# INLINE splitAt #-}
1079-- | /O(n)/ 'span', applied to a predicate @p@ and text @t@, returns
1254-- | /O(n)/ 'span', applied to a predicate @p@ and text @t@, returns
1080-- a pair whose first element is the longest prefix (possibly empty)
1255-- a pair whose first element is the longest prefix (possibly empty)
1081-- 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
1082-- remainder of the list.
1257-- remainder of the list.
1083span :: (Char -> Bool) -> Text -> (Text, Text)
1258span :: (Char -> Bool) -> Text -> (Text, Text)
1084−span 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
1085− where k = loop 0
1260+ (# hd,tl #) -> (hd,tl)
1086− loop !i | i >= len || not (p c) = i
1087− | otherwise = loop (i+d)
1088− where Iter c d = iter t i
1089{-# INLINE span #-}
1261{-# INLINE span #-}
1091-- | /O(n)/ 'break' is like 'span', but the prefix returned is
1263-- | /O(n)/ 'break' is like 'span', but the prefix returned is
@@ -1112,7 +1284,7 @@ findAIndexOrEnd q t@(Text _arr _off len) = go 0
1112 where go !i | i >= len || q c = i
1284 where go !i | i >= len || q c = i
1113 | otherwise = go (i+d)
1285 | otherwise = go (i+d)
1114 where Iter c d = iter t i
1286 where Iter c d = iter t i
1116-- | /O(n)/ Group characters in a string by equality.
1288-- | /O(n)/ Group characters in a string by equality.
1117group :: Text -> [Text]
1289group :: Text -> [Text]
1118group = groupBy (==)
1290group = groupBy (==)
@@ -1136,31 +1308,38 @@ tails t | null t = [empty]
1136-- copies to create substrings; they just construct new 'Text's that
1308-- copies to create substrings; they just construct new 'Text's that
1137-- are slices of the original.
1309-- are slices of the original.
1139−-- | /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'
1140−-- 'Text' argument, consuming the delimiter. An empty delimiter is
1312+-- argument (which cannot be empty), consuming the delimiter. An empty
1141−-- invalid, and will cause an error to be raised.
1313+-- delimiter is invalid, and will cause an error to be raised.
1143-- Examples:
1315-- Examples:
1145-- > splitOn "\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"]
1146-- > splitOn "aaa" "aaaXaaaXaaaXaaa" == ["","X","X","X",""]
1318-- > splitOn "aaa" "aaaXaaaXaaaXaaa" == ["","X","X","X",""]
1147-- > splitOn "x" "x" == ["",""]
1319-- > splitOn "x" "x" == ["",""]
1151-- > intercalate s . splitOn s == id
1323-- > intercalate s . splitOn s == id
1152-- > splitOn (singleton c) == split (==c)
1324-- > splitOn (singleton c) == split (==c)
1326+-- (Note: the string @s@ to split on above cannot be empty.)
1154-- In (unlikely) bad cases, this function's time complexity degrades
1328-- In (unlikely) bad cases, this function's time complexity degrades
1155-- towards /O(n*m)/.
1329-- towards /O(n*m)/.
1156−splitOn :: Text -> Text -> [Text]
1330+splitOn :: Text
1331+ -- ^ String to split on. If this string is empty, an error
1157splitOn pat@(Text _ _ l) src@(Text arr off len)
1336splitOn pat@(Text _ _ l) src@(Text arr off len)
1158 | l <= 0 = emptyError "splitOn"
1337 | l <= 0 = emptyError "splitOn"
1159 | isSingleton pat = split (== unsafeHead pat) src
1338 | isSingleton pat = split (== unsafeHead pat) src
1160 | otherwise = go 0 (indices pat src)
1339 | otherwise = go 0 (indices pat src)
1162− 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
1163− go s _ = [textP arr (s+off) (len-s)]
1342+ go s _ = [text arr (s+off) (len-s)]
1164{-# INLINE [1] splitOn #-}
1343{-# INLINE [1] splitOn #-}
1166{-# RULES
1345{-# RULES
@@ -1180,7 +1359,7 @@ split _ t@(Text _off _arr 0) = [t]
1180split p t = loop t
1359split p t = loop t
1181 where loop s | null s' = [l]
1360 where loop s | null s' = [l]
1182 | otherwise = l : loop (unsafeTail s')
1361 | otherwise = l : loop (unsafeTail s')
1183− where (l, s') = break p s
1362+ where (# l, s' #) = span_ (not . p) s
1184{-# INLINE split #-}
1363{-# INLINE split #-}
1186-- | /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
@@ -1252,7 +1431,7 @@ breakOn pat src@(Text arr off len)
1252 | null pat = emptyError "breakOn"
1431 | null pat = emptyError "breakOn"
1253 | otherwise = case indices pat src of
1432 | otherwise = case indices pat src of
1254 [] -> (src, empty)
1433 [] -> (src, empty)
1255− (x:_) -> (textP arr off x, textP arr (off+x) (len-x))
1434+ (x:_) -> (text arr off x, text arr (off+x) (len-x))
1256{-# INLINE breakOn #-}
1435{-# INLINE breakOn #-}
1258-- | /O(n+m)/ Similar to 'breakOn', but searches from the end of the
1437-- | /O(n+m)/ Similar to 'breakOn', but searches from the end of the
@@ -1277,9 +1456,9 @@ breakOnEnd pat src = (reverse b, reverse a)
1278-- Examples:
1457-- Examples:
1280−-- > find "::" ""
1459+-- > breakOnAll "::" ""
1281-- > ==> []
1460-- > ==> []
1282−-- > find "/" "a/b/c/"
1461+-- > breakOnAll "/" "a/b/c/"
1283-- > ==> [("a", "/b/c/"), ("a/b", "/c/"), ("a/b/c", "/")]
1462-- > ==> [("a", "/b/c/"), ("a/b", "/c/"), ("a/b/c", "/")]
1285-- In (unlikely) bad cases, this function's time complexity degrades
1464-- In (unlikely) bad cases, this function's time complexity degrades
@@ -1294,7 +1473,7 @@ breakOnAll pat src@(Text arr off slen)
1294 | otherwise = L.map step (indices pat src)
1473 | otherwise = L.map step (indices pat src)
1296 step x = (chunk 0 x, chunk x (slen-x))
1475 step x = (chunk 0 x, chunk x (slen-x))
1297− chunk !n !l = textP arr (n+off) l
1476+ chunk !n !l = text arr (n+off) l
1298{-# INLINE breakOnAll #-}
1477{-# INLINE breakOnAll #-}
1300-------------------------------------------------------------------------------
1479-------------------------------------------------------------------------------
@@ -1314,7 +1493,7 @@ breakOnAll pat src@(Text arr off slen)
1314-- 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
1315-- the substring @\"::\"@, such as @\"foo::bar::quux\"@. Instead of
1494-- the substring @\"::\"@, such as @\"foo::bar::quux\"@. Instead of
1316-- searching for the index of @\"::\"@ and taking the substrings
1495-- searching for the index of @\"::\"@ and taking the substrings
1317−-- before and after that index, you would instead use @find \"::\"@.
1496+-- before and after that index, you would instead use @breakOnAll \"::\"@.
1319-- | /O(n)/ 'Text' index (subscript) operator, starting from 0.
1498-- | /O(n)/ 'Text' index (subscript) operator, starting from 0.
1320index :: Text -> Int -> Char
1499index :: Text -> Int -> Char
@@ -1361,7 +1540,7 @@ countChar c t = S.countChar c (stream t)
1361-- equivalent to a pair of 'unpack' operations.
1540-- equivalent to a pair of 'unpack' operations.
1362zip :: Text -> Text -> [(Char,Char)]
1541zip :: Text -> Text -> [(Char,Char)]
1363zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1542zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1364−{-# INLINE [0] zip #-}
1543+{-# INLINE zip #-}
1366-- | /O(n)/ 'zipWith' generalises 'zip' by zipping with the function
1545-- | /O(n)/ 'zipWith' generalises 'zip' by zipping with the function
1367-- given as the first argument, instead of a tupling function.
1546-- given as the first argument, instead of a tupling function.
@@ -1369,7 +1548,7 @@ zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1369zipWith :: (Char -> Char -> Char) -> Text -> Text -> Text
1548zipWith :: (Char -> Char -> Char) -> Text -> Text -> Text
1370zipWith f t1 t2 = unstream (S.zipWith g (stream t1) (stream t2))
1549zipWith f t1 t2 = unstream (S.zipWith g (stream t1) (stream t2))
1371 where g a b = safe (f a b)
1550 where g a b = safe (f a b)
1372−{-# INLINE [0] zipWith #-}
1551+{-# INLINE zipWith #-}
1374-- | /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
1375-- representing white space.
1554-- representing white space.
@@ -1395,7 +1574,7 @@ lines ps | null ps = []
1395 | otherwise = h : if null t
1574 | otherwise = h : if null t
1397 else lines (unsafeTail t)
1576 else lines (unsafeTail t)
1398− where (h,t) = span (/= '\n') ps
1577+ where (# h,t #) = span_ (/= '\n') ps
1399{-# INLINE lines #-}
1578{-# INLINE lines #-}
@@ -1441,8 +1620,6 @@ isPrefixOf a@(Text _ _ alen) b@(Text _ _ blen) =
1441{-# RULES
1620{-# RULES
1442"TEXT isPrefixOf -> fused" [~1] forall s t.
1621"TEXT isPrefixOf -> fused" [~1] forall s t.
1443 isPrefixOf s t = S.isPrefixOf (stream s) (stream t)
1622 isPrefixOf s t = S.isPrefixOf (stream s) (stream t)
1444−"TEXT isPrefixOf -> unfused" [1] forall s t.
1445− S.isPrefixOf (stream s) (stream t) = isPrefixOf s t
1448-- | /O(n)/ The 'isSuffixOf' function takes two 'Text's and returns
1625-- | /O(n)/ The 'isSuffixOf' function takes two 'Text's and returns
@@ -1496,7 +1673,7 @@ isInfixOf needle haystack
1496-- > fnordLength _ = -1
1673-- > fnordLength _ = -1
1497stripPrefix :: Text -> Text -> Maybe Text
1674stripPrefix :: Text -> Text -> Maybe Text
1498stripPrefix p@(Text _arr _off plen) t@(Text arr off len)
1675stripPrefix p@(Text _arr _off plen) t@(Text arr off len)
1499− | p `isPrefixOf` t = Just $! textP arr (off+plen) (len-plen)
1676+ | p `isPrefixOf` t = Just $! text arr (off+plen) (len-plen)
1500 | otherwise = Nothing
1677 | otherwise = Nothing
1502-- | /O(n)/ Find the longest non-empty common prefix of two strings
1679-- | /O(n)/ Find the longest non-empty common prefix of two strings
@@ -1516,8 +1693,8 @@ commonPrefixes t0@(Text arr0 off0 len0) t1@(Text arr1 off1 len1) = go 0 0
1517 go !i !j | i < len0 && j < len1 && a == b = go (i+d0) (j+d1)
1694 go !i !j | i < len0 && j < len1 && a == b = go (i+d0) (j+d1)
1518 | i > 0 = Just (Text arr0 off0 i,
1695 | i > 0 = Just (Text arr0 off0 i,
1519− textP arr0 (off0+i) (len0-i),
1696+ text arr0 (off0+i) (len0-i),
1520− textP arr1 (off1+j) (len1-j))
1697+ text arr1 (off1+j) (len1-j))
1521 | otherwise = Nothing
1698 | otherwise = Nothing
1522 where Iter a d0 = iter t0 i
1699 where Iter a d0 = iter t0 i
1523 Iter b d1 = iter t1 j
1700 Iter b d1 = iter t1 j
@@ -1542,7 +1719,7 @@ commonPrefixes t0@(Text arr0 off0 len0) t1@(Text arr1 off1 len1) = go 0 0
1542-- > quuxLength _ = -1
1719-- > quuxLength _ = -1
1543stripSuffix :: Text -> Text -> Maybe Text
1720stripSuffix :: Text -> Text -> Maybe Text
1544stripSuffix p@(Text _arr _off plen) t@(Text arr off len)
1721stripSuffix p@(Text _arr _off plen) t@(Text arr off len)
1545− | p `isSuffixOf` t = Just $! textP arr off (len-plen)
1722+ | p `isSuffixOf` t = Just $! text arr off (len-plen)
1546 | otherwise = Nothing
1723 | otherwise = Nothing
1548-- | Add a list of non-negative numbers. Errors out on overflow.
1725-- | Add a list of non-negative numbers. Errors out on overflow.
@@ -1559,3 +1736,20 @@ emptyError fun = P.error $ "Data.Text." ++ fun ++ ": empty input"
1560overflowError :: String -> a
1737overflowError :: String -> a
1561overflowError 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