da754bb · typeable/attoparsec

git.iko.soy/typeable/attoparsec

Add new lookahead functions

commitda754bb3130575b482c7536a4dc51f1f7f7322bb authorBryan O'Sullivan <bos@serpentine.com> dateTue 26 Nov 2013, 21:24 (13 years ago) parenta2351e7

5 files changed+45−8

compare with master

unifiedsplit

Data/Attoparsec/ByteString.hs+4−1 Data/Attoparsec/ByteString/Char8.hs+12−3 Data/Attoparsec/ByteString/Internal.hs+13−2 Data/Attoparsec/Text.hs+3 Data/Attoparsec/Text/Internal.hs+13−2

Data/Attoparsec/ByteString.hs+4−1

@@ -47,11 +47,14 @@ module Data.Attoparsec.ByteString

4747 , I.word8

4848 , I.anyWord8

4949 , I.notWord8

50− , I.peekWord8

5150 , I.satisfy

5251 , I.satisfyWith

5352 , I.skip

5453

54+ -- ** Lookahead

55+ , I.peekWord8

56+ , I.peekWord8'

57+

5558 -- ** Byte classes

5659 , I.inClass

5760 , I.notInClass

Data/Attoparsec/ByteString/Char8.hs+12−3

@@ -46,9 +46,12 @@ module Data.Attoparsec.ByteString.Char8

4646 , char8

4747 , anyChar

4848 , notChar

49− , peekChar

5049 , satisfy

5150

51+ -- ** Lookahead

52+ , peekChar

53+ , peekChar'

54+

5255 -- ** Special character parsers

5356 , digit

5457 , letter_iso8859_15

@@ -223,8 +226,8 @@ anyChar :: Parser Char

223226anyChar = satisfy $ const True

224227{-# INLINE anyChar #-}

225228

226−-- | Match any character. Returns 'Nothing' if end of input has been

227−-- reached. Does not consume any input.

229+-- | Match any character, to perform lookahead. Returns 'Nothing' if

230+-- end of input has been reached. Does not consume any input.

228231--

229232-- /Note/: Because this parser does not fail, do not use it with

230233-- combinators such as 'many', because such parsers loop until a

@@ -233,6 +236,12 @@ peekChar :: Parser (Maybe Char)

233236peekChar = (fmap w2c) `fmap` I.peekWord8

234237{-# INLINE peekChar #-}

235238

239+-- | Match any character, to perform lookahead. Does not consume any

240+-- input, but will fail if end of input has been reached.

241+peekChar' :: Parser Char

242+peekChar' = w2c `fmap` I.peekWord8'

243+{-# INLINE peekChar' #-}

244+

236245-- | Fast predicate for matching ASCII space characters.

237246--

238247-- /Note/: This predicate only gives correct answers for the ASCII

Data/Attoparsec/ByteString/Internal.hs+13−2

@@ -34,7 +34,10 @@ module Data.Attoparsec.ByteString.Internal

3434 , skip

3535 , word8

3636 , notWord8

37+

38+ -- ** Lookahead

3739 , peekWord8

40+ , peekWord8'

3841

3942 -- ** Byte classes

4043 , inClass

@@ -426,8 +429,8 @@ notWord8 :: Word8 -> Parser Word8

426429notWord8 c = satisfy (/= c) <?> "not " ++ show c

427430{-# INLINE notWord8 #-}

428431

429−-- | Match any byte. Returns 'Nothing' if end of input has been

430−-- reached. Does not consume any input.

432+-- | Match any byte, to perform lookahead. Returns 'Nothing' if end of

433+-- input has been reached. Does not consume any input.

431434--

432435-- /Note/: Because this parser does not fail, do not use it with

433436-- combinators such as 'many', because such parsers loop until a

@@ -445,6 +448,14 @@ peekWord8 = T.Parser $ \i0 a0 m0 _kf ks ->

445448 in ks i0 a0 m0 (Just w)

446449{-# INLINE peekWord8 #-}

447450

451+-- | Match any byte, to perform lookahead. Does not consume any

452+-- input, but will fail if end of input has been reached.

453+peekWord8' :: Parser Word8

454+peekWord8' = do

455+ s <- ensure 1

456+ return $! B.unsafeHead s

457+{-# INLINE peekWord8' #-}

458+

448459-- | Match only if all input has been consumed.

449460endOfInput :: Parser ()

450461endOfInput = T.Parser $ \i0 a0 m0 kf ks ->

Data/Attoparsec/Text.hs+3

@@ -52,7 +52,10 @@ module Data.Attoparsec.Text

5252 , I.satisfy

5353 , I.satisfyWith

5454 , I.skip

55+

56+ -- ** Lookahead

5557 , I.peekChar

58+ , I.peekChar'

5659

5760 -- ** Special character parsers

5861 , digit

Data/Attoparsec/Text/Internal.hs+13−2

@@ -35,7 +35,10 @@ module Data.Attoparsec.Text.Internal

3535 , skip

3636 , char

3737 , notChar

38+

39+ -- ** Lookahead

3840 , peekChar

41+ , peekChar'

3942

4043 -- ** Character classes

4144 , inClass

@@ -440,8 +443,8 @@ notChar :: Char -> Parser Char

440443notChar c = satisfy (/= c) <?> "not " ++ show c

441444{-# INLINE notChar #-}

442445

443−-- | Match any character. Returns 'Nothing' if end of input has been

444−-- reached. Does not consume any input.

446+-- | Match any character, to perform lookahead. Returns 'Nothing' if

447+-- end of input has been reached. Does not consume any input.

445448--

446449-- /Note/: Because this parser does not fail, do not use it with

447450-- combinators such as 'many', because such parsers loop until a

@@ -459,6 +462,14 @@ peekChar = T.Parser $ \i0 a0 m0 _kf ks ->

459462 in ks i0 a0 m0 (Just c)

460463{-# INLINE peekChar #-}

461464

465+-- | Match any character, to perform lookahead. Does not consume any

466+-- input, but will fail if end of input has been reached.

467+peekChar' :: Parser Char

468+peekChar' = do

469+ s <- ensure 1

470+ return $! unsafeHead s

471+{-# INLINE peekChar' #-}

472+

462473-- | Match only if all input has been consumed.

463474endOfInput :: Parser ()

464475endOfInput = T.Parser $ \i0 a0 m0 kf ks ->