@@ -1,48 +1,29 @@
1−{-# LANGUAGE BangPatterns, CPP, MagicHash, Rank2Types, UnboxedTuples #-}
1+{-# LANGUAGE BangPatterns, CPP, Rank2Types #-}
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 #-}
11-- Module : Data.Text
5-- Module : Data.Text
6+-- Copyright : (c) Tom Harper 2008-2009,
12−-- Copyright : (c) 2009, 2010, 2011, 2012 Bryan O'Sullivan,
7+-- (c) Bryan O'Sullivan 2009,
13−-- (c) 2009 Duncan Coutts,
8+-- (c) Duncan Coutts 2009
14−-- (c) 2008, 2009 Tom Harper
16-- License : BSD-style
10-- License : BSD-style
17−-- Maintainer : bos@serpentine.com
11+-- Maintainer : bos@serpentine.com, rtharper@aftereternity.co.uk,
18-- Stability : experimental
13-- Stability : experimental
19-- Portability : GHC
14-- Portability : GHC
21−-- A time and space-efficient implementation of Unicode text.
16+-- A time and space-efficient implementation of Unicode text using
22−-- Suitable for performance critical use, both in terms of large data
17+-- packed Word16 arrays. Suitable for performance critical use, both
23−-- quantities and high speed.
18+-- in terms of large data quantities and high speed.
25−-- /Note/: Read below the synopsis for important notes on the use of
28-- This module is intended to be imported @qualified@, to avoid name
20-- This module is intended to be imported @qualified@, to avoid name
29-- clashes with "Prelude" functions, e.g.
21-- clashes with "Prelude" functions, e.g.
31-- > import qualified Data.Text as T
23-- > import qualified Data.Text as T
33−-- To use an extended and very rich family of functions for working
34−-- with Unicode text (including normalization, regular expressions,
35−-- non-standard encodings, text breaking, and locales), see
36−-- <http://hackage.haskell.org/package/text-icu the text-icu package >.
38module Data.Text
25module Data.Text
40− -- * Strict vs lazy types
46 -- * Fusion
27 -- * Fusion
47 -- $fusion
28 -- $fusion
@@ -66,7 +47,6 @@ module Data.Text
71 -- * Transformations
51 -- * Transformations
@@ -81,7 +61,6 @@ module Data.Text
81 , toCaseFold
61 , toCaseFold
86 -- ** Justification
65 -- ** Justification
87 , justifyLeft
66 , justifyLeft
@@ -125,11 +104,8 @@ module Data.Text
126 -- ** Breaking strings
105 -- ** Breaking strings
131 , takeWhile
108 , takeWhile
133 , dropWhile
109 , dropWhile
134 , dropWhileEnd
110 , dropWhileEnd
135 , dropAround
111 , dropAround
@@ -137,10 +113,10 @@ module Data.Text
137 , stripStart
113 , stripStart
138 , stripEnd
114 , stripEnd
139 , splitAt
115 , splitAt
140− , breakOn
116+ , spanBy
143− , span
118+ , breakEnd
145 , groupBy
121 , groupBy
@@ -148,8 +124,8 @@ module Data.Text
149 -- ** Breaking into many substrings
125 -- ** Breaking into many substrings
150 -- $split
126 -- $split
153 , chunksOf
129 , chunksOf
155 -- ** Breaking into lines and words
131 -- ** Breaking into lines and words
@@ -164,309 +140,156 @@ module Data.Text
164 , isSuffixOf
140 , isSuffixOf
165 , isInfixOf
141 , isInfixOf
172 -- * Searching
143 -- * Searching
176− , partition
146+ , findBy
178 -- , findSubstring
149 -- , findSubstring
180 -- * Indexing
151 -- * Indexing
181 -- $index
152 -- $index
183 , findIndex
154 , findIndex
186− -- * Zipping
157+ -- * Zipping and unzipping
188 , zipWith
159 , zipWith
190 -- -* Ordered text
161 -- -* Ordered text
191 -- , sort
162 -- , sort
193− -- * Low level operations
198−import Prelude (Char, Bool(..), Int, Maybe(..), String,
165+import Prelude (Char, Bool(..), Functor(..), Int, Maybe(..), String,
199− Eq(..), Ord(..), Ordering(..), (++),
166+ Eq(..), Ord(..), (++),
200− Read(..),
167+ Read(..), Show(..),
201− (&&), (||), (+), (-), (.), ($), ($!), (>>),
168+ (&&), (||), (+), (-), (.), ($), (>>), (*),
202− not, return, otherwise, quot)
169+ div, error, not, return, otherwise)
203#if defined(HAVE_DEEPSEQ)
170#if defined(HAVE_DEEPSEQ)
204−import Control.DeepSeq (NFData(rnf))
171+import Control.DeepSeq (NFData)
207import Control.Exception (assert)
173import Control.Exception (assert)
209import Data.Char (isSpace)
174import Data.Char (isSpace)
210−import Data.Data (Data(gfoldl, toConstr, gunfold, dataTypeOf), constrIndex,
175+import Data.Data (Data(gfoldl, toConstr, gunfold, dataTypeOf), mkNorepType)
211− Constr, mkConstr, DataType, mkDataType, Fixity(Prefix))
212import Control.Monad (foldM)
176import Control.Monad (foldM)
213import Control.Monad.ST (ST)
177import Control.Monad.ST (ST)
214import qualified Data.Text.Array as A
178import qualified Data.Text.Array as A
215import qualified Data.List as L
179import qualified Data.List as L
216−import Data.Binary (Binary(get, put))
217import Data.Monoid (Monoid(..))
180import Data.Monoid (Monoid(..))
218−#if MIN_VERSION_base(4,9,0)
219−import Data.Semigroup (Semigroup(..))
181+import Data.Word (Word16)
221import Data.String (IsString(..))
182import Data.String (IsString(..))
222−import qualified Data.Text.Internal.Fusion as S
183+import qualified Data.Text.Fusion as S
223−import qualified Data.Text.Internal.Fusion.Common as S
184+import qualified Data.Text.Fusion.Common as S
224−import Data.Text.Encoding (decodeUtf8', encodeUtf8)
225−import Data.Text.Internal.Fusion (stream, reverseStream, unstream)
185+import Data.Text.Fusion (stream, reverseStream, unstream)
226−import Data.Text.Internal.Private (span_)
186+import Data.Text.Internal (Text(..), empty, text, textP)
227−import Data.Text.Internal (Text(..), empty, firstf, mul, safe, text)
228−import Data.Text.Show (singleton, unpack, unpackCString#)
229import qualified Prelude as P
187import qualified Prelude as P
230−import Data.Text.Unsafe (Iter(..), iter, iter_, lengthWord16, reverseIter,
188+import Data.Text.Unsafe (iter, iter_, reverseIter, unsafeHead, unsafeTail)
231− reverseIter_, unsafeHead, unsafeTail)
232−import Data.Text.Internal.Unsafe.Char (unsafeChr)
189+import Data.Text.UnsafeChar (unsafeChr)
233−import qualified Data.Text.Internal.Functions as F
190+import qualified Data.Text.Encoding.Utf16 as U16
234−import qualified Data.Text.Internal.Encoding.Utf16 as U16
235−import Data.Text.Internal.Search (indices)
191+import Data.Text.Search (indices)
236−#if defined(__HADDOCK__)
237−import Data.ByteString (ByteString)
238−import qualified Data.Text.Lazy as L
239−import 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)
251−-- This package provides both strict and lazy 'Text' types. The
252−-- strict type is provided by the "Data.Text" module, while the lazy
253−-- type is provided by the "Data.Text.Lazy" module. Internally, the
254−-- lazy @Text@ type consists of a list of strict chunks.
256−-- The strict 'Text' type requires that an entire string fit into
257−-- memory at once. The lazy 'Data.Text.Lazy.Text' type is capable of
258−-- streaming strings that are larger than memory using a small memory
259−-- footprint. In many cases, the overhead of chunked streaming makes
260−-- the lazy 'Data.Text.Lazy.Text' type slower than its strict
261−-- counterpart, but this is not always the case. Sometimes, the time
262−-- complexity of a function in one module may be different from the
263−-- other, due to their differing internal structures.
265−-- Each module provides an almost identical API, with the main
266−-- difference being that the strict module uses 'Int' values for
267−-- lengths and counts, while the lazy module uses 'Data.Int.Int64'
272−-- A 'Text' value is a sequence of Unicode scalar values, as defined
274−-- <http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=35 §3.9, definition D76 of the Unicode 5.2 standard >.
275−-- As such, a 'Text' cannot contain values in the range U+D800 to
276−-- U+DFFF inclusive. Haskell implementations admit all Unicode code
278−-- (<http://www.unicode.org/versions/Unicode5.2.0/ch03.pdf#page=13 §3.4, definition D10 >)
279−-- as 'Char' values, including code points from this invalid range.
280−-- This means that there are some 'Char' values that are not valid
281−-- Unicode scalar values, and the functions in this module must handle
284−-- Within this module, many functions construct a 'Text' from one or
285−-- more 'Char' values. Those functions will substitute 'Char' values
286−-- that are not valid Unicode scalar values with the replacement
287−-- character \"�\" (U+FFFD). Functions that perform this
288−-- inspection and replacement are documented with the phrase
289−-- \"Performs replacement on invalid scalar values\".
291−-- (One reason for this policy of replacement is that internally, a
292−-- 'Text' value is represented as packed UTF-16 data. Values in the
293−-- range U+D800 through U+DFFF are used by UTF-16 to denote surrogate
294−-- code points, and so cannot be represented. The functions replace
295−-- invalid scalar values, instead of dropping them, as a security
296−-- measure. For details, see
297−-- <http://unicode.org/reports/tr36/#Deletion_of_Noncharacters Unicode Technical Report 36, §3.5 >.)
299-- $fusion
193-- $fusion
301-- Most of the functions in this module are subject to /fusion/,
195-- Most of the functions in this module are subject to /fusion/,
302-- meaning that a pipeline of such functions will usually allocate at
196-- meaning that a pipeline of such functions will usually allocate at
303-- most one 'Text' value.
197-- most one 'Text' value.
305−-- As an example, consider the following pipeline:
307−-- > import Data.Text as T
308−-- > import Data.Text.Encoding as E
309−-- > import Data.ByteString (ByteString)
311−-- > countChars :: ByteString -> Int
312−-- > countChars = T.length . T.toUpper . E.decodeUtf8
314−-- From the type signatures involved, this looks like it should
315−-- allocate one 'Data.ByteString.ByteString' value, and two 'Text'
316−-- values. However, when a module is compiled with optimisation
317−-- enabled under GHC, the two intermediate 'Text' values will be
318−-- optimised away, and the function will be compiled down to a single
319−-- loop over the source 'Data.ByteString.ByteString'.
321−-- Functions that can be fused by the compiler are documented with the
322−-- phrase \"Subject to fusion\".
324instance Eq Text where
199instance Eq Text where
325− Text arrA offA lenA == Text arrB offB lenB
200+ t1 == t2 = stream t1 == stream t2
326− | lenA == lenB = A.equal arrA offA arrB offB lenA
328 {-# INLINE (==) #-}
201 {-# INLINE (==) #-}
330instance Ord Text where
203instance Ord Text where
331− compare = compareText
204+ compare t1 t2 = compare (stream t1) (stream t2)
205+ {-# INLINE compare #-}
207+instance Show Text where
208+ showsPrec p ps r = showsPrec p (unpack ps) r
333instance Read Text where
210instance Read Text where
334 readsPrec p str = [(pack x,y) | (x,y) <- readsPrec p str]
211 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
344instance Monoid Text where
213instance Monoid Text where
345 mempty = empty
214 mempty = empty
346−#if MIN_VERSION_base(4,9,0)
347− mappend = (<>) -- future-proof definition
349 mappend = append
215 mappend = append
351 mconcat = concat
216 mconcat = concat
353instance IsString Text where
218instance IsString Text where
354 fromString = pack
219 fromString = pack
356−#if __GLASGOW_HASKELL__ >= 708
357−instance Exts.IsList Text where
358− type Item Text = Char
363#if defined(HAVE_DEEPSEQ)
221#if defined(HAVE_DEEPSEQ)
364−instance NFData Text where rnf !_ = ()
222+instance NFData Text
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
375−-- | This instance preserves data abstraction at the cost of inefficiency.
225+-- This instance preserves data abstraction at the cost of inefficiency.
376-- We omit reflection services for the sake of data abstraction.
226-- We omit reflection services for the sake of data abstraction.
228+-- This instance was created by copying the behavior of Data.Set and
229+-- Data.Map. If you feel a mistake has been made, please feel free to
230+-- submit improvements.
378−-- This instance was created by copying the updated behavior of
379−-- @"Data.Set".@'Data.Set.Set' and @"Data.Map".@'Data.Map.Map'. If you
380−-- feel a mistake has been made, please feel free to submit
383−-- The original discussion is archived here:
384−-- <http://groups.google.com/group/haskell-cafe/browse_thread/thread/b5bbb1b28a7e525d/0639d46852575b93 could we get a Data instance for Data.Text.Text? >
386−-- The followup discussion that changed the behavior of 'Data.Set.Set'
387−-- and 'Data.Map.Map' is archived here:
232+-- Original discussion is archived here:
388−-- <http://markmail.org/message/trovdc6zkphyi3cr#query:+page:1+mid:a46der3iacwjcf6n+state:results Proposal: Allow gunfold for Data.Map, ... >
234+-- "could we get a Data instance for Data.Text.Text?"
235+-- http://groups.google.com/group/haskell-cafe/browse_thread/thread/b5bbb1b28a7e525d/0639d46852575b93
237+-- NOTE: we used the mispelled version of mkNoRepType for
238+-- compatibility with GHC 6.10. Eventually this should be updated:
239+-- http://hackage.haskell.org/trac/ghc/ticket/2760
390instance Data Text where
240instance Data Text where
391 gfoldl f z txt = z pack `f` (unpack txt)
241 gfoldl f z txt = z pack `f` (unpack txt)
392− toConstr _ = packConstr
242+ toConstr _ = error "Data.Text.Text:toConstr"
393− gunfold k z c = case constrIndex c of
395− _ -> P.error "gunfold"
243+ gunfold _ _ = error "Data.Text.Text:gunfold"
396− dataTypeOf _ = textDataType
244+ dataTypeOf _ = mkNorepType "Data.Text.Text"
398−#if MIN_VERSION_base(4,7,0)
399−-- | Only defined for @base-4.7.0.0@ and later
400−instance PrintfArg Text where
401− formatArg txt = formatString $ unpack txt
405−packConstr = mkConstr textDataType "pack" [] Prefix
407−textDataType :: DataType
408−textDataType = mkDataType "Data.Text.Text" [packConstr]
410−-- | /O(n)/ Compare two 'Text' values lexicographically.
411−compareText :: Text -> Text -> Ordering
412−compareText ta@(Text _arrA _offA lenA) tb@(Text _arrB _offB lenB)
413− | lenA == 0 && lenB == 0 = EQ
414− | otherwise = go 0 0
417− | i >= lenA || j >= lenB = compare lenA lenB
420− | otherwise = go (i+di) (j+dj)
421− where Iter a di = iter ta i
422− Iter b dj = iter tb j
424-- -----------------------------------------------------------------------------
246-- -----------------------------------------------------------------------------
425-- * Conversion to/from 'Text'
247-- * Conversion to/from 'Text'
427−-- | /O(n)/ Convert a 'String' into a 'Text'. Subject to
249+-- | /O(n)/ Convert a 'String' into a 'Text'. Subject to fusion.
428−-- fusion. Performs replacement on invalid scalar values.
429pack :: String -> Text
250pack :: String -> Text
430−pack = unstream . S.map safe . S.streamList
251+pack = unstream . S.streamList
431{-# INLINE [1] pack #-}
252{-# INLINE [1] pack #-}
254+-- | /O(n)/ Convert a Text into a String. Subject to fusion.
255+unpack :: Text -> String
256+unpack = S.unstreamList . stream
257+{-# INLINE [1] unpack #-}
259+-- | /O(1)/ Convert a character into a Text.
260+-- Subject to fusion.
261+singleton :: Char -> Text
262+singleton = unstream . S.singleton
263+{-# INLINE [1] singleton #-}
433-- -----------------------------------------------------------------------------
265-- -----------------------------------------------------------------------------
434-- * Basic functions
266-- * Basic functions
436-- | /O(n)/ Adds a character to the front of a 'Text'. This function
268-- | /O(n)/ Adds a character to the front of a 'Text'. This function
437-- is more costly than its 'List' counterpart because it requires
269-- is more costly than its 'List' counterpart because it requires
438−-- copying a new array. Subject to fusion. Performs replacement on
270+-- copying a new array. Subject to fusion.
439−-- invalid scalar values.
440cons :: Char -> Text -> Text
271cons :: Char -> Text -> Text
441−cons c t = unstream (S.cons (safe c) (stream t))
272+cons c t = unstream (S.cons c (stream t))
442{-# INLINE cons #-}
273{-# INLINE cons #-}
446-- | /O(n)/ Adds a character to the end of a 'Text'. This copies the
275-- | /O(n)/ Adds a character to the end of a 'Text'. This copies the
447-- entire array in the process, unless fused. Subject to fusion.
276-- entire array in the process, unless fused. Subject to fusion.
448−-- Performs replacement on invalid scalar values.
449snoc :: Text -> Char -> Text
277snoc :: Text -> Char -> Text
450−snoc t c = unstream (S.snoc (stream t) (safe c))
278+snoc t c = unstream (S.snoc (stream t) c)
451{-# INLINE snoc #-}
279{-# INLINE snoc #-}
453-- | /O(n)/ Appends one 'Text' to the other by copying both of them
281-- | /O(n)/ Appends one 'Text' to the other by copying both of them
454-- into a new 'Text'. Subject to fusion.
282-- into a new 'Text'. Subject to fusion.
455append :: Text -> Text -> Text
283append :: Text -> Text -> Text
456−append a@(Text arr1 off1 len1) b@(Text arr2 off2 len2)
284+append (Text arr1 off1 len1) (Text arr2 off2 len2) = Text (A.run x) 0 len
459− | len > 0 = Text (A.run x) 0 len
460− | otherwise = overflowError "append"
462 len = len1+len2
286 len = len1+len2
463− x :: ST s (A.MArray s)
465− arr <- A.new len
288+ arr <- A.unsafeNew len
466− A.copyI arr 0 arr1 off1 len1
289+ copy arr 0 arr1 off1 len1
467− A.copyI arr len1 arr2 off2 len
290+ copy arr len1 arr2 off2 (len1+len2)
468 return arr
291 return arr
469−{-# NOINLINE append #-}
292+{-# INLINE append #-}
472"TEXT append -> fused" [~1] forall t1 t2.
295"TEXT append -> fused" [~1] forall t1 t2.
@@ -475,6 +298,14 @@ append a@(Text arr1 off1 len1) b@(Text arr2 off2 len2)
475 unstream (S.append (stream t1) (stream t2)) = append t1 t2
298 unstream (S.append (stream t1) (stream t2)) = append t1 t2
301+copy :: forall s. A.MArray s Word16 -> Int -> A.Array Word16 -> Int -> Int
303+copy dest i0 src j0 top = go i0 j0
305+ go i j | i >= top = return ()
306+ | otherwise = do A.unsafeWrite dest i (src `A.unsafeIndex` j)
478-- | /O(1)/ Returns the first character of a 'Text', which must be
309-- | /O(1)/ Returns the first character of a 'Text', which must be
479-- non-empty. Subject to fusion.
310-- non-empty. Subject to fusion.
480head :: Text -> Char
311head :: Text -> Char
@@ -486,14 +317,21 @@ head t = S.head (stream t)
486uncons :: Text -> Maybe (Char, Text)
317uncons :: Text -> Maybe (Char, Text)
487uncons t@(Text arr off len)
318uncons t@(Text arr off len)
488 | len <= 0 = Nothing
319 | len <= 0 = Nothing
489− | otherwise = Just $ let !(Iter c d) = iter t 0
490− in (c, text arr (off+d) (len-d))
320+ | otherwise = Just (c, textP arr (off+d) (len-d))
321+ where (c,d) = iter t 0
491{-# INLINE [1] uncons #-}
322{-# INLINE [1] uncons #-}
493-- | Lifted from Control.Arrow and specialized.
324-- | Lifted from Control.Arrow and specialized.
494second :: (b -> c) -> (a,b) -> (a,c)
325second :: (b -> c) -> (a,b) -> (a,c)
495second f (a, b) = (a, f b)
326second f (a, b) = (a, f b)
329+"TEXT uncons -> fused" [~1] forall t.
330+ uncons t = fmap (second unstream) (S.uncons (stream t))
331+"TEXT uncons -> unfused" [1] forall t.
332+ fmap (second unstream) (S.uncons (stream t)) = uncons t
497-- | /O(1)/ Returns the last character of a 'Text', which must be
335-- | /O(1)/ Returns the last character of a 'Text', which must be
498-- non-empty. Subject to fusion.
336-- non-empty. Subject to fusion.
499last :: Text -> Char
337last :: Text -> Char
@@ -517,7 +355,7 @@ last (Text arr off len)
517tail :: Text -> Text
355tail :: Text -> Text
518tail t@(Text arr off len)
356tail t@(Text arr off len)
519 | len <= 0 = emptyError "tail"
357 | len <= 0 = emptyError "tail"
520− | otherwise = text arr (off+d) (len-d)
358+ | otherwise = textP arr (off+d) (len-d)
521 where d = iter_ t 0
359 where d = iter_ t 0
522{-# INLINE [1] tail #-}
360{-# INLINE [1] tail #-}
@@ -532,8 +370,8 @@ tail t@(Text arr off len)
532-- be non-empty. Subject to fusion.
370-- be non-empty. Subject to fusion.
533init :: Text -> Text
371init :: Text -> Text
534init (Text arr off len) | len <= 0 = emptyError "init"
372init (Text arr off len) | len <= 0 = emptyError "init"
535− | n >= 0xDC00 && n <= 0xDFFF = text arr off (len-2)
373+ | n >= 0xDC00 && n <= 0xDFFF = textP arr off (len-2)
536− | otherwise = text arr off (len-1)
374+ | otherwise = textP arr off (len-1)
538 n = A.unsafeIndex arr (off+len-1)
376 n = A.unsafeIndex arr (off+len-1)
539{-# INLINE [1] init #-}
377{-# INLINE [1] init #-}
@@ -548,11 +386,7 @@ init (Text arr off len) | len <= 0 = emptyError "init"
548-- | /O(1)/ Tests whether a 'Text' is empty or not. Subject to
386-- | /O(1)/ Tests whether a 'Text' is empty or not. Subject to
549-- fusion.
387-- fusion.
550null :: Text -> Bool
388null :: Text -> Bool
551−null (Text _arr _off len) =
389+null (Text _arr _off len) = assert (len >= 0) $ len <= 0
556{-# INLINE [1] null #-}
390{-# INLINE [1] null #-}
@@ -572,76 +406,27 @@ isSingleton = S.isSingleton . stream
572-- Subject to fusion.
406-- Subject to fusion.
573length :: Text -> Int
407length :: Text -> Int
574length t = S.length (stream t)
408length t = S.length (stream t)
575−{-# INLINE [0] length #-}
409+{-# INLINE 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.
579−-- | /O(n)/ Compare the count of characters in a 'Text' to a number.
580−-- Subject to fusion.
582−-- This function gives the same answer as comparing against the result
583−-- of 'length', but can short circuit if the count of characters is
584−-- greater than the number, and hence be more efficient.
585−compareLength :: Text -> Int -> Ordering
586−compareLength t n = S.compareLengthI (stream t) n
587−{-# INLINE [1] compareLength #-}
590−"TEXT compareN/length -> compareLength" [~1] forall t n.
591− compare (length t) n = compareLength t n
595−"TEXT ==N/length -> compareLength/==EQ" [~1] forall t n.
596− eqInt (length t) n = compareLength t n == EQ
600−"TEXT /=N/length -> compareLength//=EQ" [~1] forall t n.
601− neInt (length t) n = compareLength t n /= EQ
605−"TEXT <N/length -> compareLength/==LT" [~1] forall t n.
606− ltInt (length t) n = compareLength t n == LT
610−"TEXT <=N/length -> compareLength//=GT" [~1] forall t n.
611− leInt (length t) n = compareLength t n /= GT
615−"TEXT >N/length -> compareLength/==GT" [~1] forall t n.
616− gtInt (length t) n = compareLength t n == GT
620−"TEXT >=N/length -> compareLength//=LT" [~1] forall t n.
621− geInt (length t) n = compareLength t n /= LT
624-- -----------------------------------------------------------------------------
411-- -----------------------------------------------------------------------------
625-- * Transformations
412-- * Transformations
626-- | /O(n)/ 'map' @f@ @t@ is the 'Text' obtained by applying @f@ to
413-- | /O(n)/ 'map' @f@ @t@ is the 'Text' obtained by applying @f@ to
627−-- each element of @t@. Subject to fusion. Performs replacement on
414+-- each element of @t@. Subject to fusion.
628−-- invalid scalar values.
629map :: (Char -> Char) -> Text -> Text
415map :: (Char -> Char) -> Text -> Text
630−map f t = unstream (S.map (safe . f) (stream t))
416+map f t = unstream (S.map f (stream t))
631{-# INLINE [1] map #-}
417{-# INLINE [1] map #-}
633-- | /O(n)/ The 'intercalate' function takes a 'Text' and a list of
419-- | /O(n)/ The 'intercalate' function takes a 'Text' and a list of
634-- 'Text's and concatenates the list after interspersing the first
420-- 'Text's and concatenates the list after interspersing the first
635-- argument between each element of the list.
421-- argument between each element of the list.
636intercalate :: Text -> [Text] -> Text
422intercalate :: Text -> [Text] -> Text
637−intercalate t = concat . (F.intersperse t)
423+intercalate t = concat . (L.intersperse t)
638{-# INLINE intercalate #-}
424{-# INLINE intercalate #-}
640-- | /O(n)/ The 'intersperse' function takes a character and places it
426-- | /O(n)/ The 'intersperse' function takes a character and places it
641−-- between the characters of a 'Text'. Subject to fusion. Performs
427+-- between the characters of a 'Text'. Subject to fusion.
642−-- replacement on invalid scalar values.
643intersperse :: Char -> Text -> Text
428intersperse :: Char -> Text -> Text
644−intersperse c t = unstream (S.intersperse (safe c) (stream t))
429+intersperse c t = unstream (S.intersperse c (stream t))
645{-# INLINE intersperse #-}
430{-# INLINE intersperse #-}
647-- | /O(n)/ Reverse the characters of a string. Subject to fusion.
432-- | /O(n)/ Reverse the characters of a string. Subject to fusion.
@@ -649,59 +434,13 @@ reverse :: Text -> Text
649reverse t = S.reverse (stream t)
434reverse t = S.reverse (stream t)
650{-# INLINE reverse #-}
435{-# INLINE reverse #-}
652−-- | /O(m+n)/ Replace every non-overlapping occurrence of @needle@ in
437+-- | /O(m+n)/ Replace every occurrence of one substring with another.
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"
673−-- In (unlikely) bad cases, this function's time complexity degrades
675−replace :: Text
438+replace :: Text -- ^ Text to search for
676− -- ^ @needle@ to search for. If this string is empty, an
439+ -> Text -- ^ Replacement text
677− -- error will occur.
440+ -> Text -- ^ Input text
679− -- ^ @replacement@ to replace @needle@ with.
442+replace s d = intercalate d . split s
680− -> Text
443+{-# INLINE replace #-}
681− -- ^ @haystack@ in which to search.
683−replace needle@(Text _ _ neeLen)
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
706-- ----------------------------------------------------------------------------
445-- ----------------------------------------------------------------------------
707-- ** Case conversions (folds)
446-- ** Case conversions (folds)
@@ -716,17 +455,10 @@ replace needle@(Text _ _ neeLen)
716-- case conversion rules. As a result, these functions may map one
455-- case conversion rules. As a result, these functions may map one
717-- input character to two or three output characters. For examples,
456-- input character to two or three output characters. For examples,
718-- see the documentation of each function.
457-- see the documentation of each function.
720−-- /Note/: In some languages, case conversion is a locale- and
721−-- context-dependent operation. The case conversion functions in this
722−-- module are /not/ locale sensitive. Programs that require locale
723−-- sensitivity should use appropriate versions of the
724−-- <http://hackage.haskell.org/package/text-icu-0.6.3.7/docs/Data-Text-ICU.html#g:4 case mapping functions from the text-icu package >.
726−-- | /O(n)/ Convert a string to folded case. Subject to fusion.
459+-- | /O(n)/ Convert a string to folded case. This function is mainly
728−-- This function is mainly useful for performing caseless (also known
460+-- useful for performing caseless (also known as case insensitive)
729−-- as case insensitive) string comparisons.
461+-- string comparisons.
731-- A string @x@ is a caseless match for a string @y@ if and only if:
463-- A string @x@ is a caseless match for a string @y@ if and only if:
@@ -741,56 +473,27 @@ replace needle@(Text _ _ neeLen)
741-- instead of itself.
473-- instead of itself.
742toCaseFold :: Text -> Text
474toCaseFold :: Text -> Text
743toCaseFold t = unstream (S.toCaseFold (stream t))
475toCaseFold t = unstream (S.toCaseFold (stream t))
744−{-# INLINE toCaseFold #-}
476+{-# INLINE [0] toCaseFold #-}
746-- | /O(n)/ Convert a string to lower case, using simple case
478-- | /O(n)/ Convert a string to lower case, using simple case
747−-- conversion. Subject to fusion.
749−-- The result string may be longer than the input string. For
479+-- conversion. The result string may be longer than the input string.
750−-- instance, \"İ\" (Latin capital letter I with dot above,
480+-- For instance, \"İ\" (Latin capital letter I with dot above,
751−-- U+0130) maps to the sequence \"i\" (Latin small letter i, U+0069)
481+-- U+0130) maps to the sequence \"i\" (Latin small letter i, U+0069) followed
752−-- followed by \" ̇\" (combining dot above, U+0307).
482+-- by \" ̇\" (combining dot above, U+0307).
753toLower :: Text -> Text
483toLower :: Text -> Text
754toLower t = unstream (S.toLower (stream t))
484toLower t = unstream (S.toLower (stream t))
755{-# INLINE toLower #-}
485{-# INLINE toLower #-}
757-- | /O(n)/ Convert a string to upper case, using simple case
487-- | /O(n)/ Convert a string to upper case, using simple case
758−-- conversion. Subject to fusion.
760−-- The result string may be longer than the input string. For
488+-- conversion. The result string may be longer than the input string.
761−-- instance, the German \"ß\" (eszett, U+00DF) maps to the
489+-- For instance, the German \"ß\" (eszett, U+00DF) maps to the
762-- two-letter sequence \"SS\".
490-- two-letter sequence \"SS\".
763toUpper :: Text -> Text
491toUpper :: Text -> Text
764toUpper t = unstream (S.toUpper (stream t))
492toUpper t = unstream (S.toUpper (stream t))
765{-# INLINE toUpper #-}
493{-# 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 #-}
789-- | /O(n)/ Left-justify a string to the given length, using the
495-- | /O(n)/ Left-justify a string to the given length, using the
790−-- specified fill character on the right. Subject to fusion.
496+-- specified fill character on the right. Subject to fusion. Examples:
791−-- Performs replacement on invalid scalar values.
795-- > justifyLeft 7 'x' "foo" == "fooxxxx"
498-- > justifyLeft 7 'x' "foo" == "fooxxxx"
796-- > justifyLeft 3 'x' "foobar" == "foobar"
499-- > justifyLeft 3 'x' "foobar" == "foobar"
@@ -809,10 +512,7 @@ justifyLeft k c t
811-- | /O(n)/ Right-justify a string to the given length, using the
514-- | /O(n)/ Right-justify a string to the given length, using the
812−-- specified fill character on the left. Performs replacement on
515+-- specified fill character on the left. Examples:
813−-- invalid scalar values.
817-- > justifyRight 7 'x' "bar" == "xxxxbar"
517-- > justifyRight 7 'x' "bar" == "xxxxbar"
818-- > justifyRight 3 'x' "foobar" == "foobar"
518-- > justifyRight 3 'x' "foobar" == "foobar"
@@ -823,11 +523,8 @@ justifyRight k c t
823 where len = length t
523 where len = length t
824{-# INLINE justifyRight #-}
524{-# INLINE justifyRight #-}
826−-- | /O(n)/ Center a string to the given length, using the specified
526+-- | /O(n)/ Center a string to the given length, using the
827−-- fill character on either side. Performs replacement on invalid
527+-- specified fill character on either side. Examples:
832-- > center 8 'x' "HS" = "xxxHSxxx"
529-- > center 8 'x' "HS" = "xxxHSxxx"
833center :: Int -> Char -> Text -> Text
530center :: Int -> Char -> Text -> Text
@@ -836,7 +533,7 @@ center k c t
836 | otherwise = replicateChar l c `append` t `append` replicateChar r c
533 | otherwise = replicateChar l c `append` t `append` replicateChar r c
837 where len = length t
534 where len = length t
838 d = k - len
535 d = k - len
839− r = d `quot` 2
536+ r = d `div` 2
840 l = d - r
537 l = d - r
841{-# INLINE center #-}
538{-# INLINE center #-}
@@ -854,12 +551,12 @@ transpose ts = P.map pack (L.transpose (P.map unpack ts))
854-- (typically the left-identity of the operator), and a 'Text',
551-- (typically the left-identity of the operator), and a 'Text',
855-- reduces the 'Text' using the binary operator, from left to right.
552-- reduces the 'Text' using the binary operator, from left to right.
856-- Subject to fusion.
553-- Subject to fusion.
857−foldl :: (a -> Char -> a) -> a -> Text -> a
554+foldl :: (b -> Char -> b) -> b -> Text -> b
858foldl f z t = S.foldl f z (stream t)
555foldl f z t = S.foldl f z (stream t)
859{-# INLINE foldl #-}
556{-# INLINE foldl #-}
861-- | /O(n)/ A strict version of 'foldl'. Subject to fusion.
558-- | /O(n)/ A strict version of 'foldl'. Subject to fusion.
862−foldl' :: (a -> Char -> a) -> a -> Text -> a
559+foldl' :: (b -> Char -> b) -> b -> Text -> b
863foldl' f z t = S.foldl' f z (stream t)
560foldl' f z t = S.foldl' f z (stream t)
864{-# INLINE foldl' #-}
561{-# INLINE foldl' #-}
@@ -878,12 +575,12 @@ foldl1' f t = S.foldl1' f (stream t)
878-- (typically the right-identity of the operator), and a 'Text',
575-- (typically the right-identity of the operator), and a 'Text',
879-- reduces the 'Text' using the binary operator, from right to left.
576-- reduces the 'Text' using the binary operator, from right to left.
880-- Subject to fusion.
577-- Subject to fusion.
881−foldr :: (Char -> a -> a) -> a -> Text -> a
578+foldr :: (Char -> b -> b) -> b -> Text -> b
882foldr f z t = S.foldr f z (stream t)
579foldr f z t = S.foldr f z (stream t)
883{-# INLINE foldr #-}
580{-# INLINE foldr #-}
885-- | /O(n)/ A variant of 'foldr' that has no starting value argument,
582-- | /O(n)/ A variant of 'foldr' that has no starting value argument,
886−-- and thus must be applied to a non-empty 'Text'. Subject to
583+-- and thust must be applied to a non-empty 'Text'. Subject to
887-- fusion.
584-- fusion.
888foldr1 :: (Char -> Char -> Char) -> Text -> Char
585foldr1 :: (Char -> Char -> Char) -> Text -> Char
889foldr1 f t = S.foldr1 f (stream t)
586foldr1 f t = S.foldr1 f (stream t)
@@ -894,19 +591,14 @@ foldr1 f t = S.foldr1 f (stream t)
895-- | /O(n)/ Concatenate a list of 'Text's.
592-- | /O(n)/ Concatenate a list of 'Text's.
896concat :: [Text] -> Text
593concat :: [Text] -> Text
897−concat ts = case ts' of
900− _ -> Text (A.run go) 0 len
594+concat ts = Text (A.run go) 0 len
902− ts' = L.filter (not . null) ts
596+ len = L.sum (L.map (\(Text _ _ l) -> l) ts)
903− len = sumP "concat" $ L.map lengthWord16 ts'
904− go :: ST s (A.MArray s)
906− arr <- A.new len
598+ arr <- A.unsafeNew len
907− let step i (Text a o l) =
908− let !j = i + l in A.copyI arr i a o j >> return j
599+ let step i (Text a o l) = let j = i + l in copy arr i a o j >> return j
909− foldM step 0 ts' >> return arr
600+ foldM step 0 ts >> return arr
601+{-# INLINE concat #-}
911-- | /O(n)/ Map a function over a 'Text' that results in a 'Text', and
603-- | /O(n)/ Map a function over a 'Text' that results in a 'Text', and
912-- concatenate the results.
604-- concatenate the results.
@@ -915,13 +607,13 @@ concatMap f = concat . foldr ((:) . f) []
915{-# INLINE concatMap #-}
607{-# INLINE concatMap #-}
917-- | /O(n)/ 'any' @p@ @t@ determines whether any character in the
609-- | /O(n)/ 'any' @p@ @t@ determines whether any character in the
918−-- 'Text' @t@ satisfies the predicate @p@. Subject to fusion.
610+-- 'Text' @t@ satisifes the predicate @p@. Subject to fusion.
919any :: (Char -> Bool) -> Text -> Bool
611any :: (Char -> Bool) -> Text -> Bool
920any p t = S.any p (stream t)
612any p t = S.any p (stream t)
921{-# INLINE any #-}
613{-# INLINE any #-}
923-- | /O(n)/ 'all' @p@ @t@ determines whether all characters in the
615-- | /O(n)/ 'all' @p@ @t@ determines whether all characters in the
924−-- 'Text' @t@ satisfy the predicate @p@. Subject to fusion.
616+-- 'Text' @t@ satisify the predicate @p@. Subject to fusion.
925all :: (Char -> Bool) -> Text -> Bool
617all :: (Char -> Bool) -> Text -> Bool
926all p t = S.all p (stream t)
618all p t = S.all p (stream t)
927{-# INLINE all #-}
619{-# INLINE all #-}
@@ -943,7 +635,6 @@ minimum t = S.minimum (stream t)
944-- | /O(n)/ 'scanl' is similar to 'foldl', but returns a list of
636-- | /O(n)/ 'scanl' is similar to 'foldl', but returns a list of
945-- successive reduced values from the left. Subject to fusion.
637-- successive reduced values from the left. Subject to fusion.
946−-- Performs replacement on invalid scalar values.
948-- > scanl f z [x1, x2, ...] == [z, z `f` x1, (z `f` x1) `f` x2, ...]
639-- > scanl f z [x1, x2, ...] == [z, z `f` x1, (z `f` x1) `f` x2, ...]
@@ -951,13 +642,11 @@ minimum t = S.minimum (stream t)
952-- > last (scanl f z xs) == foldl f z xs.
643-- > last (scanl f z xs) == foldl f z xs.
953scanl :: (Char -> Char -> Char) -> Char -> Text -> Text
644scanl :: (Char -> Char -> Char) -> Char -> Text -> Text
954−scanl f z t = unstream (S.scanl g z (stream t))
645+scanl f z t = unstream (S.scanl f z (stream t))
955− where g a b = safe (f a b)
956{-# INLINE scanl #-}
646{-# INLINE scanl #-}
958-- | /O(n)/ 'scanl1' is a variant of 'scanl' that has no starting
648-- | /O(n)/ 'scanl1' is a variant of 'scanl' that has no starting
959−-- value argument. Subject to fusion. Performs replacement on
649+-- value argument. Subject to fusion.
960−-- invalid scalar values.
962-- > scanl1 f [x1, x2, ...] == [x1, x1 `f` x2, ...]
651-- > scanl1 f [x1, x2, ...] == [x1, x1 `f` x2, ...]
963scanl1 :: (Char -> Char -> Char) -> Text -> Text
652scanl1 :: (Char -> Char -> Char) -> Text -> Text
@@ -965,42 +654,40 @@ scanl1 f t | null t = empty
965 | otherwise = scanl f (unsafeHead t) (unsafeTail t)
654 | otherwise = scanl f (unsafeHead t) (unsafeTail t)
966{-# INLINE scanl1 #-}
655{-# INLINE scanl1 #-}
968−-- | /O(n)/ 'scanr' is the right-to-left dual of 'scanl'. Performs
657+-- | /O(n)/ 'scanr' is the right-to-left dual of 'scanl'.
969−-- replacement on invalid scalar values.
971-- > scanr f v == reverse . scanl (flip f) v . reverse
659-- > scanr f v == reverse . scanl (flip f) v . reverse
972scanr :: (Char -> Char -> Char) -> Char -> Text -> Text
660scanr :: (Char -> Char -> Char) -> Char -> Text -> Text
973−scanr f z = S.reverse . S.reverseScanr g z . reverseStream
661+scanr f z = S.reverse . S.reverseScanr f z . reverseStream
974− where g a b = safe (f a b)
975{-# INLINE scanr #-}
662{-# INLINE scanr #-}
977-- | /O(n)/ 'scanr1' is a variant of 'scanr' that has no starting
664-- | /O(n)/ 'scanr1' is a variant of 'scanr' that has no starting
978−-- value argument. Subject to fusion. Performs replacement on
665+-- value argument. Subject to fusion.
979−-- invalid scalar values.
980scanr1 :: (Char -> Char -> Char) -> Text -> Text
666scanr1 :: (Char -> Char -> Char) -> Text -> Text
981scanr1 f t | null t = empty
667scanr1 f t | null t = empty
982 | otherwise = scanr f (last t) (init t)
668 | otherwise = scanr f (last t) (init t)
983{-# INLINE scanr1 #-}
669{-# INLINE scanr1 #-}
985−-- | /O(n)/ Like a combination of 'map' and 'foldl''. Applies a
671+-- | /O(n)/ Like a combination of 'map' and 'foldl'. Applies a
986-- function to each element of a 'Text', passing an accumulating
672-- function to each element of a 'Text', passing an accumulating
987−-- parameter from left to right, and returns a final 'Text'. Performs
673+-- parameter from left to right, and returns a final 'Text'.
988−-- replacement on invalid scalar values.
989mapAccumL :: (a -> Char -> (a,Char)) -> a -> Text -> (a, Text)
674mapAccumL :: (a -> Char -> (a,Char)) -> a -> Text -> (a, Text)
990−mapAccumL f z0 = S.mapAccumL g z0 . stream
675+mapAccumL f s t = case uncons t of
991− where g a b = second safe (f a b)
676+ Nothing -> (s, empty)
677+ Just (x, xs) -> (s'', cons y ys)
678+ where (s', y ) = f s x
992−{-# INLINE mapAccumL #-}
679+ (s'',ys) = mapAccumL f s' xs
994-- | The 'mapAccumR' function behaves like a combination of 'map' and
681-- | The 'mapAccumR' function behaves like a combination of 'map' and
995−-- a strict 'foldr'; it applies a function to each element of a
682+-- 'foldr'; it applies a function to each element of a 'Text', passing
996−-- 'Text', passing an accumulating parameter from right to left, and
683+-- an accumulating parameter from right to left, and returning a final
997−-- returning a final value of this accumulator together with the new
684+-- value of this accumulator together with the new 'Text'.
999−-- Performs replacement on invalid scalar values.
1000mapAccumR :: (a -> Char -> (a,Char)) -> a -> Text -> (a, Text)
685mapAccumR :: (a -> Char -> (a,Char)) -> a -> Text -> (a, Text)
1001−mapAccumR f z0 = second reverse . S.mapAccumL g z0 . reverseStream
686+mapAccumR f s t = case uncons t of
1002− where g a b = second safe (f a b)
687+ Nothing -> (s, empty)
688+ Just (x, xs) -> (s'', cons y ys)
689+ where (s'',y ) = f s' x
1003−{-# INLINE mapAccumR #-}
690+ (s', ys) = mapAccumR f s xs
1005-- -----------------------------------------------------------------------------
692-- -----------------------------------------------------------------------------
1006-- ** Generating and unfolding 'Text's
693-- ** Generating and unfolding 'Text's
@@ -1009,18 +696,17 @@ mapAccumR f z0 = second reverse . S.mapAccumL g z0 . reverseStream
1009-- @t@ repeated @n@ times.
696-- @t@ repeated @n@ times.
1010replicate :: Int -> Text -> Text
697replicate :: Int -> Text -> Text
1011replicate n t@(Text a o l)
698replicate n t@(Text a o l)
1012− | n <= 0 || l <= 0 = empty
699+ | n <= 0 || l <= 0 = empty
1013− | n == 1 = t
700+ | n == 1 = t
1014− | isSingleton t = replicateChar n (unsafeHead t)
701+ | isSingleton t = replicateChar n (unsafeHead t)
1015− | otherwise = Text (A.run x) 0 len
702+ | otherwise = Text (A.run x) 0 len
1017− len = l `mul` n
704+ len = l * n
1018− x :: ST s (A.MArray s)
1020− arr <- A.new len
706+ arr <- A.unsafeNew len
1021 let loop !d !i | i >= n = return arr
707 let loop !d !i | i >= n = return arr
1022 | otherwise = let m = d + l
708 | otherwise = let m = d + l
1023− in A.copyI arr d a o m >> loop m (i+1)
709+ in copy arr d a o m >> loop m (i+1)
1024 loop 0 0
710 loop 0 0
1025{-# INLINE [1] replicate #-}
711{-# INLINE [1] replicate #-}
@@ -1032,7 +718,7 @@ replicate n t@(Text a o l)
1032-- | /O(n)/ 'replicateChar' @n@ @c@ is a 'Text' of length @n@ with @c@ the
718-- | /O(n)/ 'replicateChar' @n@ @c@ is a 'Text' of length @n@ with @c@ the
1033-- value of every element. Subject to fusion.
719-- value of every element. Subject to fusion.
1034replicateChar :: Int -> Char -> Text
720replicateChar :: Int -> Char -> Text
1035−replicateChar n c = unstream (S.replicateCharI n (safe c))
721+replicateChar n c = unstream (S.replicateCharI n c)
1036{-# INLINE replicateChar #-}
722{-# INLINE replicateChar #-}
1038-- | /O(n)/, where @n@ is the length of the result. The 'unfoldr'
724-- | /O(n)/, where @n@ is the length of the result. The 'unfoldr'
@@ -1041,9 +727,9 @@ replicateChar n c = unstream (S.replicateCharI n (safe c))
1041-- returns 'Nothing' if it is done producing the 'Text', otherwise
727-- returns 'Nothing' if it is done producing the 'Text', otherwise
1042-- 'Just' @(a,b)@. In this case, @a@ is the next 'Char' in the
728-- 'Just' @(a,b)@. In this case, @a@ is the next 'Char' in the
1043-- string, and @b@ is the seed value for further production. Subject
729-- string, and @b@ is the seed value for further production. Subject
1044−-- to fusion. Performs replacement on invalid scalar values.
730+-- to fusion.
1045unfoldr :: (a -> Maybe (Char,a)) -> a -> Text
731unfoldr :: (a -> Maybe (Char,a)) -> a -> Text
1046−unfoldr f s = unstream (S.unfoldr (firstf safe . f) s)
732+unfoldr f s = unstream (S.unfoldr f s)
1047{-# INLINE unfoldr #-}
733{-# INLINE unfoldr #-}
1049-- | /O(n)/ Like 'unfoldr', 'unfoldrN' builds a 'Text' from a seed
735-- | /O(n)/ Like 'unfoldr', 'unfoldrN' builds a 'Text' from a seed
@@ -1051,9 +737,9 @@ unfoldr f s = unstream (S.unfoldr (firstf safe . f) s)
1051-- first argument to 'unfoldrN'. This function is more efficient than
737-- first argument to 'unfoldrN'. This function is more efficient than
1052-- 'unfoldr' when the maximum length of the result is known and
738-- 'unfoldr' when the maximum length of the result is known and
1053-- correct, otherwise its performance is similar to 'unfoldr'. Subject
739-- correct, otherwise its performance is similar to 'unfoldr'. Subject
1054−-- to fusion. Performs replacement on invalid scalar values.
740+-- to fusion.
1055unfoldrN :: Int -> (a -> Maybe (Char,a)) -> a -> Text
741unfoldrN :: Int -> (a -> Maybe (Char,a)) -> a -> Text
1056−unfoldrN n f s = unstream (S.unfoldrN n (firstf safe . f) s)
742+unfoldrN n f s = unstream (S.unfoldrN n f s)
1057{-# INLINE unfoldrN #-}
743{-# INLINE unfoldrN #-}
1059-- -----------------------------------------------------------------------------
745-- -----------------------------------------------------------------------------
@@ -1066,16 +752,14 @@ take :: Int -> Text -> Text
1066take n t@(Text arr off len)
752take n t@(Text arr off len)
1067 | n <= 0 = empty
753 | n <= 0 = empty
1068 | n >= len = t
754 | n >= len = t
1069− | otherwise = text arr off (iterN n t)
755+ | otherwise = Text arr off (loop 0 0)
758+ | i >= len || cnt >= n = i
759+ | otherwise = loop (i+d) (cnt+1)
1070{-# INLINE [1] take #-}
761{-# 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
1079{-# RULES
763{-# RULES
1080"TEXT take -> fused" [~1] forall n t.
764"TEXT take -> fused" [~1] forall n t.
1081 take n t = unstream (S.take n (stream t))
765 take n t = unstream (S.take n (stream t))
@@ -1083,36 +767,18 @@ iterN n t@(Text _arr _off len) = loop 0 0
1083 unstream (S.take n (stream t)) = take n t
767 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
1107-- | /O(n)/ 'drop' @n@, applied to a 'Text', returns the suffix of the
770-- | /O(n)/ 'drop' @n@, applied to a 'Text', returns the suffix of the
1108−-- 'Text' after the first @n@ characters, or the empty 'Text' if @n@
771+-- 'Text' of length @n@, or the empty 'Text' if @n@ is greater than the
1109−-- is greater than the length of the 'Text'. Subject to fusion.
772+-- length of the 'Text'. Subject to fusion.
1110drop :: Int -> Text -> Text
773drop :: Int -> Text -> Text
1111drop n t@(Text arr off len)
774drop n t@(Text arr off len)
1112 | n <= 0 = t
775 | n <= 0 = t
1113 | n >= len = empty
776 | n >= len = empty
1114− | otherwise = text arr (off+i) (len-i)
777+ | otherwise = loop 0 0
1115− where i = iterN n t
778+ where loop !i !cnt
779+ | i >= len || cnt >= n = Text arr (off+i) (len-i)
780+ | otherwise = loop (i+d) (cnt+1)
1116{-# INLINE [1] drop #-}
782{-# INLINE [1] drop #-}
1118{-# RULES
784{-# RULES
@@ -1122,18 +788,6 @@ drop n t@(Text arr off len)
1122 unstream (S.drop n (stream t)) = drop n t
788 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)
1137-- | /O(n)/ 'takeWhile', applied to a predicate @p@ and a 'Text',
791-- | /O(n)/ 'takeWhile', applied to a predicate @p@ and a 'Text',
1138-- returns the longest prefix (possibly empty) of elements that
792-- returns the longest prefix (possibly empty) of elements that
1139-- satisfy @p@. Subject to fusion.
793-- satisfy @p@. Subject to fusion.
@@ -1141,8 +795,8 @@ takeWhile :: (Char -> Bool) -> Text -> Text
1141takeWhile p t@(Text arr off len) = loop 0
795takeWhile p t@(Text arr off len) = loop 0
1142 where loop !i | i >= len = t
796 where loop !i | i >= len = t
1143 | p c = loop (i+d)
797 | p c = loop (i+d)
1144− | otherwise = text arr off i
798+ | otherwise = textP arr off i
1145− where Iter c d = iter t i
799+ where (c,d) = iter t i
1146{-# INLINE [1] takeWhile #-}
800{-# INLINE [1] takeWhile #-}
1148{-# RULES
802{-# RULES
@@ -1152,27 +806,6 @@ takeWhile p t@(Text arr off len) = loop 0
1152 unstream (S.takeWhile p (stream t)) = takeWhile p t
806 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
1176-- | /O(n)/ 'dropWhile' @p@ @t@ returns the suffix remaining after
809-- | /O(n)/ 'dropWhile' @p@ @t@ returns the suffix remaining after
1177-- 'takeWhile' @p@ @t@. Subject to fusion.
810-- 'takeWhile' @p@ @t@. Subject to fusion.
1178dropWhile :: (Char -> Bool) -> Text -> Text
811dropWhile :: (Char -> Bool) -> Text -> Text
@@ -1180,7 +813,7 @@ dropWhile p t@(Text arr off len) = loop 0 0
1180 where loop !i !l | l >= len = empty
813 where loop !i !l | l >= len = empty
1181 | p c = loop (i+d) (l+d)
814 | p c = loop (i+d) (l+d)
1182 | otherwise = Text arr (off+i) (len-l)
815 | otherwise = Text arr (off+i) (len-l)
1183− where Iter c d = iter t i
816+ where (c,d) = iter t i
1184{-# INLINE [1] dropWhile #-}
817{-# INLINE [1] dropWhile #-}
1186{-# RULES
819{-# RULES
@@ -1191,9 +824,8 @@ dropWhile p t@(Text arr off len) = loop 0 0
1193-- | /O(n)/ 'dropWhileEnd' @p@ @t@ returns the prefix remaining after
826-- | /O(n)/ 'dropWhileEnd' @p@ @t@ returns the prefix remaining after
1194−-- dropping characters that satisfy the predicate @p@ from the end of
827+-- dropping characters that fail the predicate @p@ from the end of
1195-- @t@. Subject to fusion.
828-- @t@. Subject to fusion.
1197-- Examples:
829-- Examples:
1199-- > dropWhileEnd (=='.') "foo..." == "foo"
831-- > dropWhileEnd (=='.') "foo..." == "foo"
@@ -1213,7 +845,7 @@ dropWhileEnd p t@(Text arr off len) = loop (len-1) len
1215-- | /O(n)/ 'dropAround' @p@ @t@ returns the substring remaining after
847-- | /O(n)/ 'dropAround' @p@ @t@ returns the substring remaining after
1216−-- dropping characters that satisfy the predicate @p@ from both the
848+-- dropping characters that fail the predicate @p@ from both the
1217-- beginning and end of @t@. Subject to fusion.
849-- beginning and end of @t@. Subject to fusion.
1218dropAround :: (Char -> Bool) -> Text -> Text
850dropAround :: (Char -> Bool) -> Text -> Text
1219dropAround p = dropWhile p . dropWhileEnd p
851dropAround p = dropWhile p . dropWhileEnd p
@@ -1248,23 +880,31 @@ splitAt :: Int -> Text -> (Text, Text)
1248splitAt n t@(Text arr off len)
880splitAt n t@(Text arr off len)
1249 | n <= 0 = (empty, t)
881 | n <= 0 = (empty, t)
1250 | n >= len = (t, empty)
882 | n >= len = (t, empty)
1251− | otherwise = let k = iterN n t
883+ | otherwise = (Text arr off k, Text arr (off+k) (len-k))
1252− in (text arr off k, text arr (off+k) (len-k))
884+ where k = loop 0 0
886+ | i >= len || cnt >= n = i
887+ | otherwise = loop (i+d) (cnt+1)
889+{-# INLINE splitAt #-}
1254−-- | /O(n)/ 'span', applied to a predicate @p@ and text @t@, returns
891+-- | /O(n)/ 'spanBy', applied to a predicate @p@ and text @t@, returns
1255-- a pair whose first element is the longest prefix (possibly empty)
892-- a pair whose first element is the longest prefix (possibly empty)
1256-- of @t@ of elements that satisfy @p@, and whose second is the
893-- of @t@ of elements that satisfy @p@, and whose second is the
1257-- remainder of the list.
894-- remainder of the list.
1258−span :: (Char -> Bool) -> Text -> (Text, Text)
895+spanBy :: (Char -> Bool) -> Text -> (Text, Text)
1259−span p t = case span_ p t of
896+spanBy p t@(Text arr off len) = (textP arr off k, textP arr (off+k) (len-k))
1260− (# hd,tl #) -> (hd,tl)
897+ where k = loop 0
898+ loop !i | i >= len || not (p c) = i
899+ | otherwise = loop (i+d)
900+ where (c,d) = iter t i
1261−{-# INLINE span #-}
901+{-# INLINE spanBy #-}
1263−-- | /O(n)/ 'break' is like 'span', but the prefix returned is
903+-- | /O(n)/ 'breakBy' is like 'spanBy', but the prefix returned is
1264-- over elements that fail the predicate @p@.
904-- over elements that fail the predicate @p@.
1265−break :: (Char -> Bool) -> Text -> (Text, Text)
905+breakBy :: (Char -> Bool) -> Text -> (Text, Text)
1266−break p = span (not . p)
906+breakBy p = spanBy (not . p)
1267−{-# INLINE break #-}
907+{-# INLINE breakBy #-}
1269-- | /O(n)/ Group characters in a string according to a predicate.
909-- | /O(n)/ Group characters in a string according to a predicate.
1270groupBy :: (Char -> Char -> Bool) -> Text -> [Text]
910groupBy :: (Char -> Char -> Bool) -> Text -> [Text]
@@ -1273,7 +913,7 @@ groupBy p = loop
1273 loop t@(Text arr off len)
913 loop t@(Text arr off len)
1274 | null t = []
914 | null t = []
1275 | otherwise = text arr off n : loop (text arr (off+n) (len-n))
915 | otherwise = text arr off n : loop (text arr (off+n) (len-n))
1276− where Iter c d = iter t 0
916+ where (c,d) = iter t 0
1277 n = d + findAIndexOrEnd (not . p c) (Text arr (off+d) (len-d))
917 n = d + findAIndexOrEnd (not . p c) (Text arr (off+d) (len-d))
1279-- | Returns the /array/ index (in units of 'Word16') at which a
919-- | Returns the /array/ index (in units of 'Word16') at which a
@@ -1283,8 +923,8 @@ findAIndexOrEnd :: (Char -> Bool) -> Text -> Int
1283findAIndexOrEnd q t@(Text _arr _off len) = go 0
923findAIndexOrEnd q t@(Text _arr _off len) = go 0
1284 where go !i | i >= len || q c = i
924 where go !i | i >= len || q c = i
1285 | otherwise = go (i+d)
925 | otherwise = go (i+d)
1286− where Iter c d = iter t i
926+ where (c,d) = iter t i
1288-- | /O(n)/ Group characters in a string by equality.
928-- | /O(n)/ Group characters in a string by equality.
1289group :: Text -> [Text]
929group :: Text -> [Text]
1290group = groupBy (==)
930group = groupBy (==)
@@ -1308,43 +948,36 @@ tails t | null t = [empty]
1308-- copies to create substrings; they just construct new 'Text's that
948-- copies to create substrings; they just construct new 'Text's that
1309-- are slices of the original.
949-- are slices of the original.
1311−-- | /O(m+n)/ Break a 'Text' into pieces separated by the first 'Text'
951+-- | /O(m+n)/ Break a 'Text' into pieces separated by the first
1312−-- argument (which cannot be empty), consuming the delimiter. An empty
952+-- 'Text' argument, consuming the delimiter. An empty delimiter is
1313−-- delimiter is invalid, and will cause an error to be raised.
953+-- invalid, and will cause an error to be raised.
1315-- Examples:
955-- Examples:
1317−-- > splitOn "\r\n" "a\r\nb\r\nd\r\ne" == ["a","b","d","e"]
957+-- > split "\r\n" "a\r\nb\r\nd\r\ne" == ["a","b","d","e"]
1318−-- > splitOn "aaa" "aaaXaaaXaaaXaaa" == ["","X","X","X",""]
958+-- > split "aaa" "aaaXaaaXaaaXaaa" == ["","X","X","X",""]
1319−-- > splitOn "x" "x" == ["",""]
959+-- > split "x" "x" == ["",""]
1323−-- > intercalate s . splitOn s == id
963+-- > intercalate s . split s == id
1324−-- > splitOn (singleton c) == split (==c)
964+-- > split (singleton c) == splitBy (==c)
1326−-- (Note: the string @s@ to split on above cannot be empty.)
1328-- In (unlikely) bad cases, this function's time complexity degrades
966-- In (unlikely) bad cases, this function's time complexity degrades
1329-- towards /O(n*m)/.
967-- towards /O(n*m)/.
1330−splitOn :: Text
968+split :: Text -> Text -> [Text]
1331− -- ^ String to split on. If this string is empty, an error
1336−splitOn pat@(Text _ _ l) src@(Text arr off len)
969+split pat@(Text _ _ l) src@(Text arr off len)
1337− | l <= 0 = emptyError "splitOn"
970+ | l <= 0 = emptyError "split"
1338− | isSingleton pat = split (== unsafeHead pat) src
971+ | isSingleton pat = splitBy (== unsafeHead pat) src
1339 | otherwise = go 0 (indices pat src)
972 | otherwise = go 0 (indices pat src)
1341− go !s (x:xs) = text arr (s+off) (x-s) : go (x+l) xs
974+ go !s (x:xs) = textP arr (s+off) (x-s) : go (x+l) xs
1342− go s _ = [text arr (s+off) (len-s)]
975+ go s _ = [textP arr (s+off) (len-s)]
1343−{-# INLINE [1] splitOn #-}
976+{-# INLINE [1] split #-}
1345{-# RULES
978{-# RULES
1346−"TEXT splitOn/singleton -> split/==" [~1] forall c t.
979+"TEXT split/singleton -> splitBy/==" [~1] forall c t.
1347− splitOn (singleton c) t = split (==c) t
980+ split (singleton c) t = splitBy (==c) t
1350-- | /O(n)/ Splits a 'Text' into components delimited by separators,
983-- | /O(n)/ Splits a 'Text' into components delimited by separators,
@@ -1352,15 +985,15 @@ splitOn pat@(Text _ _ l) src@(Text arr off len)
1352-- resulting components do not contain the separators. Two adjacent
985-- resulting components do not contain the separators. Two adjacent
1353-- separators result in an empty component in the output. eg.
986-- separators result in an empty component in the output. eg.
1355−-- > split (=='a') "aabbaca" == ["","","bb","c",""]
988+-- > splitBy (=='a') "aabbaca" == ["","","bb","c",""]
1356−-- > split (=='a') "" == [""]
989+-- > splitBy (=='a') "" == [""]
1357−split :: (Char -> Bool) -> Text -> [Text]
990+splitBy :: (Char -> Bool) -> Text -> [Text]
1358−split _ t@(Text _off _arr 0) = [t]
991+splitBy _ t@(Text _off _arr 0) = [t]
1359−split p t = loop t
992+splitBy p t = loop t
1360 where loop s | null s' = [l]
993 where loop s | null s' = [l]
1361 | otherwise = l : loop (unsafeTail s')
994 | otherwise = l : loop (unsafeTail s')
1362− where (# l, s' #) = span_ (not . p) s
995+ where (l, s') = breakBy p s
1363−{-# INLINE split #-}
996+{-# INLINE splitBy #-}
1365-- | /O(n)/ Splits a 'Text' into components of length @k@. The last
998-- | /O(n)/ Splits a 'Text' into components of length @k@. The last
1366-- element may be shorter than the other chunks, depending on the
999-- element may be shorter than the other chunks, depending on the
@@ -1382,21 +1015,21 @@ chunksOf k = go
1382-------------------------------------------------------------------------------
1015-------------------------------------------------------------------------------
1383-- ** Searching with a predicate
1016-- ** Searching with a predicate
1385−-- | /O(n)/ The 'find' function takes a predicate and a 'Text', and
1018+-- | /O(n)/ The 'findBy' function takes a predicate and a 'Text', and
1386−-- returns the first element matching the predicate, or 'Nothing' if
1019+-- returns the first element in matching the predicate, or 'Nothing'
1387−-- there is no such element.
1020+-- if there is no such element.
1388−find :: (Char -> Bool) -> Text -> Maybe Char
1021+findBy :: (Char -> Bool) -> Text -> Maybe Char
1389−find p t = S.findBy p (stream t)
1022+findBy p t = S.findBy p (stream t)
1390−{-# INLINE find #-}
1023+{-# INLINE findBy #-}
1392−-- | /O(n)/ The 'partition' function takes a predicate and a 'Text',
1025+-- | /O(n)/ The 'partitionBy' function takes a predicate and a 'Text',
1393-- and returns the pair of 'Text's with elements which do and do not
1026-- and returns the pair of 'Text's with elements which do and do not
1394-- satisfy the predicate, respectively; i.e.
1027-- satisfy the predicate, respectively; i.e.
1396−-- > partition p t == (filter p t, filter (not . p) t)
1029+-- > partitionBy p t == (filter p t, filter (not . p) t)
1397−partition :: (Char -> Bool) -> Text -> (Text, Text)
1030+partitionBy :: (Char -> Bool) -> Text -> (Text, Text)
1398−partition p t = (filter p t, filter (not . p) t)
1031+partitionBy p t = (filter p t, filter (not . p) t)
1399−{-# INLINE partition #-}
1032+{-# INLINE partitionBy #-}
1401-- | /O(n)/ 'filter', applied to a predicate and a 'Text',
1034-- | /O(n)/ 'filter', applied to a predicate and a 'Text',
1402-- returns a 'Text' containing those characters that satisfy the
1035-- returns a 'Text' containing those characters that satisfy the
@@ -1412,69 +1045,71 @@ filter p t = unstream (S.filter p (stream t))
1413-- Examples:
1046-- Examples:
1415−-- > breakOn "::" "a::b::c" ==> ("a", "::b::c")
1048+-- > break "::" "a::b::c" ==> ("a", "::b::c")
1416−-- > breakOn "/" "foobar" ==> ("foobar", "")
1049+-- > break "/" "foobar" ==> ("foobar", "")
1420-- > append prefix match == haystack
1053-- > append prefix match == haystack
1421−-- > where (prefix, match) = breakOn needle haystack
1054+-- > where (prefix, match) = break needle haystack
1423-- If you need to break a string by a substring repeatedly (e.g. you
1056-- If you need to break a string by a substring repeatedly (e.g. you
1424−-- want to break on every instance of a substring), use 'breakOnAll'
1057+-- want to break on every instance of a substring), use 'find'
1425-- instead, as it has lower startup overhead.
1058-- instead, as it has lower startup overhead.
1427-- In (unlikely) bad cases, this function's time complexity degrades
1060-- In (unlikely) bad cases, this function's time complexity degrades
1428-- towards /O(n*m)/.
1061-- towards /O(n*m)/.
1429−breakOn :: Text -> Text -> (Text, Text)
1062+break :: Text -> Text -> (Text, Text)
1430−breakOn pat src@(Text arr off len)
1063+break pat src@(Text arr off len)
1431− | null pat = emptyError "breakOn"
1064+ | null pat = emptyError "break"
1432 | otherwise = case indices pat src of
1065 | otherwise = case indices pat src of
1433 [] -> (src, empty)
1066 [] -> (src, empty)
1434− (x:_) -> (text arr off x, text arr (off+x) (len-x))
1067+ (x:_) -> (textP arr off x, textP arr (off+x) (len-x))
1435−{-# INLINE breakOn #-}
1068+{-# INLINE break #-}
1437−-- | /O(n+m)/ Similar to 'breakOn', but searches from the end of the
1070+-- | /O(n+m)/ Similar to 'break', but searches from the end of the string.
1440-- The first element of the returned tuple is the prefix of @haystack@
1072-- The first element of the returned tuple is the prefix of @haystack@
1441-- up to and including the last match of @needle@. The second is the
1073-- up to and including the last match of @needle@. The second is the
1442-- remainder of @haystack@, following the match.
1074-- remainder of @haystack@, following the match.
1444−-- > breakOnEnd "::" "a::b::c" ==> ("a::b::", "c")
1076+-- > breakEnd "::" "a::b::c" ==> ("a::b::", "c")
1445−breakOnEnd :: Text -> Text -> (Text, Text)
1077+breakEnd :: Text -> Text -> (Text, Text)
1446−breakOnEnd pat src = (reverse b, reverse a)
1447− where (a,b) = breakOn (reverse pat) (reverse src)
1078+breakEnd pat src = let (a,b) = break (reverse pat) (reverse src)
1079+ in (reverse b, reverse a)
1448−{-# INLINE breakOnEnd #-}
1080+{-# INLINE breakEnd #-}
1450-- | /O(n+m)/ Find all non-overlapping instances of @needle@ in
1082-- | /O(n+m)/ Find all non-overlapping instances of @needle@ in
1451−-- @haystack@. Each element of the returned list consists of a pair:
1083+-- @haystack@. The first element of the returned pair is the prefix
1452−--
1084+-- of @haystack@ prior to any matches of @needle@. The second is a
1453−-- * The entire string prior to the /k/th match (i.e. the prefix)
1085+-- list of pairs.
1455−-- * The /k/th match, followed by the remainder of the string
1087+-- The first element of each pair in the list is a span from the
1088+-- beginning of a match to the beginning of the next match, while the
1089+-- second is a span from the beginning of the match to the end of the
1457-- Examples:
1092-- Examples:
1459−-- > breakOnAll "::" ""
1094+-- > find "::" ""
1460−-- > ==> []
1095+-- > ==> ("", [])
1461−-- > breakOnAll "/" "a/b/c/"
1096+-- > find "/" "a/b/c/d"
1462−-- > ==> [("a", "/b/c/"), ("a/b", "/c/"), ("a/b/c", "/")]
1097+-- > ==> ("a", [("/b","/b/c/d"), ("/c","/c/d"), ("/d","/d")])
1464-- In (unlikely) bad cases, this function's time complexity degrades
1099-- In (unlikely) bad cases, this function's time complexity degrades
1465-- towards /O(n*m)/.
1100-- towards /O(n*m)/.
1467−-- The @needle@ parameter may not be empty.
1468−breakOnAll :: Text -- ^ @needle@ to search for
1469− -> Text -- ^ @haystack@ in which to search
1470− -> [(Text, Text)]
1101+find :: Text -> Text -> (Text, [(Text, Text)])
1471−breakOnAll pat src@(Text arr off slen)
1102+find pat src@(Text arr off len)
1472− | null pat = emptyError "breakOnAll"
1103+ | null pat = emptyError "find"
1473− | otherwise = L.map step (indices pat src)
1104+ | otherwise = case indices pat src of
1106+ (x:xs) -> (chunk 0 x, go x xs)
1475− step x = (chunk 0 x, chunk x (slen-x))
1108+ go !s (x:xs) = (chunk s (x-s), chunk s (len-s)) : go x xs
1109+ go s _ = let c = chunk s (len-s)
1476− chunk !n !l = text arr (n+off) l
1111+ chunk !n !l = textP arr (n+off) l
1477−{-# INLINE breakOnAll #-}
1112+{-# INLINE find #-}
1479-------------------------------------------------------------------------------
1114-------------------------------------------------------------------------------
1480-- ** Indexing 'Text's
1115-- ** Indexing 'Text's
@@ -1493,7 +1128,7 @@ breakOnAll pat src@(Text arr off slen)
1493-- For example, suppose you have a string that you want to split on
1128-- For example, suppose you have a string that you want to split on
1494-- the substring @\"::\"@, such as @\"foo::bar::quux\"@. Instead of
1129-- the substring @\"::\"@, such as @\"foo::bar::quux\"@. Instead of
1495-- searching for the index of @\"::\"@ and taking the substrings
1130-- searching for the index of @\"::\"@ and taking the substrings
1496−-- before and after that index, you would instead use @breakOnAll \"::\"@.
1131+-- before and after that index, you would instead use @find \"::\"@.
1498-- | /O(n)/ 'Text' index (subscript) operator, starting from 0.
1133-- | /O(n)/ 'Text' index (subscript) operator, starting from 0.
1499index :: Text -> Int -> Char
1134index :: Text -> Int -> Char
@@ -1540,15 +1175,13 @@ countChar c t = S.countChar c (stream t)
1540-- equivalent to a pair of 'unpack' operations.
1175-- equivalent to a pair of 'unpack' operations.
1541zip :: Text -> Text -> [(Char,Char)]
1176zip :: Text -> Text -> [(Char,Char)]
1542zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1177zip a b = S.unstreamList $ S.zipWith (,) (stream a) (stream b)
1543−{-# INLINE zip #-}
1178+{-# INLINE [0] zip #-}
1545-- | /O(n)/ 'zipWith' generalises 'zip' by zipping with the function
1180-- | /O(n)/ 'zipWith' generalises 'zip' by zipping with the function
1546-- given as the first argument, instead of a tupling function.
1181-- given as the first argument, instead of a tupling function.
1547−-- Performs replacement on invalid scalar values.
1548zipWith :: (Char -> Char -> Char) -> Text -> Text -> Text
1182zipWith :: (Char -> Char -> Char) -> Text -> Text -> Text
1549−zipWith f t1 t2 = unstream (S.zipWith g (stream t1) (stream t2))
1183+zipWith f t1 t2 = unstream (S.zipWith f (stream t1) (stream t2))
1550− where g a b = safe (f a b)
1551−{-# INLINE zipWith #-}
1184+{-# INLINE [0] zipWith #-}
1553-- | /O(n)/ Breaks a 'Text' up into a list of words, delimited by 'Char's
1186-- | /O(n)/ Breaks a 'Text' up into a list of words, delimited by 'Char's
1554-- representing white space.
1187-- representing white space.
@@ -1564,7 +1197,7 @@ words t@(Text arr off len) = loop 0 0
1564 then loop (start+1) (start+1)
1197 then loop (start+1) (start+1)
1565 else Text arr (start+off) (n-start) : loop (n+d) (n+d)
1198 else Text arr (start+off) (n-start) : loop (n+d) (n+d)
1566 | otherwise = loop start (n+d)
1199 | otherwise = loop start (n+d)
1567− where Iter c d = iter t n
1200+ where (c,d) = iter t n
1568{-# INLINE words #-}
1201{-# INLINE words #-}
1570-- | /O(n)/ Breaks a 'Text' up into a list of 'Text's at
1203-- | /O(n)/ Breaks a 'Text' up into a list of 'Text's at
@@ -1574,7 +1207,7 @@ lines ps | null ps = []
1574 | otherwise = h : if null t
1207 | otherwise = h : if null t
1576 else lines (unsafeTail t)
1209 else lines (unsafeTail t)
1577− where (# h,t #) = span_ (/= '\n') ps
1210+ where (h,t) = spanBy (/= '\n') ps
1578{-# INLINE lines #-}
1211{-# INLINE lines #-}
@@ -1611,7 +1244,8 @@ unwords = intercalate (singleton ' ')
1611{-# INLINE unwords #-}
1244{-# INLINE unwords #-}
1613-- | /O(n)/ The 'isPrefixOf' function takes two 'Text's and returns
1246-- | /O(n)/ The 'isPrefixOf' function takes two 'Text's and returns
1614−-- 'True' iff the first is a prefix of the second. Subject to fusion.
1247+-- 'True' iff the first is a prefix of the second. This function is
1248+-- subject to fusion.
1615isPrefixOf :: Text -> Text -> Bool
1249isPrefixOf :: Text -> Text -> Bool
1616isPrefixOf a@(Text _ _ alen) b@(Text _ _ blen) =
1250isPrefixOf a@(Text _ _ alen) b@(Text _ _ blen) =
1617 alen <= blen && S.isPrefixOf (stream a) (stream b)
1251 alen <= blen && S.isPrefixOf (stream a) (stream b)
@@ -1620,6 +1254,8 @@ isPrefixOf a@(Text _ _ alen) b@(Text _ _ blen) =
1620{-# RULES
1254{-# RULES
1621"TEXT isPrefixOf -> fused" [~1] forall s t.
1255"TEXT isPrefixOf -> fused" [~1] forall s t.
1622 isPrefixOf s t = S.isPrefixOf (stream s) (stream t)
1256 isPrefixOf s t = S.isPrefixOf (stream s) (stream t)
1257+"TEXT isPrefixOf -> unfused" [1] forall s t.
1258+ S.isPrefixOf (stream s) (stream t) = isPrefixOf s t
1625-- | /O(n)/ The 'isSuffixOf' function takes two 'Text's and returns
1261-- | /O(n)/ The 'isSuffixOf' function takes two 'Text's and returns
@@ -1650,106 +1286,5 @@ isInfixOf needle haystack
1650 isInfixOf (singleton n) h = S.elem n (S.stream h)
1286 isInfixOf (singleton n) h = S.elem n (S.stream h)
1653−-------------------------------------------------------------------------------
1656−-- | /O(n)/ Return the suffix of the second string if its prefix
1657−-- matches the entire first string.
1661−-- > stripPrefix "foo" "foobar" == Just "bar"
1662−-- > stripPrefix "" "baz" == Just "baz"
1663−-- > stripPrefix "foo" "quux" == Nothing
1665−-- This is particularly useful with the @ViewPatterns@ extension to
1668−-- > {-# LANGUAGE ViewPatterns #-}
1669−-- > import Data.Text as T
1671−-- > fnordLength :: Text -> Int
1672−-- > fnordLength (stripPrefix "fnord" -> Just suf) = T.length suf
1673−-- > fnordLength _ = -1
1674−stripPrefix :: Text -> Text -> Maybe Text
1675−stripPrefix p@(Text _arr _off plen) t@(Text arr off len)
1676− | p `isPrefixOf` t = Just $! text arr (off+plen) (len-plen)
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
1702−-- | /O(n)/ Return the prefix of the second string if its suffix
1703−-- matches the entire first string.
1707−-- > stripSuffix "bar" "foobar" == Just "foo"
1708−-- > stripSuffix "" "baz" == Just "baz"
1709−-- > stripSuffix "foo" "quux" == Nothing
1711−-- This is particularly useful with the @ViewPatterns@ extension to
1714−-- > {-# LANGUAGE ViewPatterns #-}
1715−-- > import Data.Text as T
1717−-- > quuxLength :: Text -> Int
1718−-- > quuxLength (stripSuffix "quux" -> Just pre) = T.length pre
1719−-- > quuxLength _ = -1
1720−stripSuffix :: Text -> Text -> Maybe Text
1721−stripSuffix p@(Text _arr _off plen) t@(Text arr off len)
1722− | p `isSuffixOf` t = Just $! text arr off (len-plen)
1723− | otherwise = Nothing
1725−-- | Add a list of non-negative numbers. Errors out on overflow.
1726−sumP :: String -> [Int] -> Int
1729− | ax >= 0 = go ax xs
1730− | otherwise = overflowError fun
1734emptyError :: String -> a
1289emptyError :: String -> a
1735−emptyError fun = P.error $ "Data.Text." ++ fun ++ ": empty input"
1290+emptyError fun = P.error ("Data.Text." ++ fun ++ ": empty input")
1737−overflowError :: String -> a
1738−overflowError 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