bytesmith 0.3.8.0 → 0.3.9.0

raw patch · 8 files changed

+127/−61 lines, 8 filesdep ~basedep ~byteslicedep ~bytestringPVP: major bump suggested

API removals or changes: PVP suggests a major version bump

Dependency ranges changed: base, byteslice, bytestring, contiguous

API changes (from Hackage documentation)

Files

• CHANGELOG.md +4/−0
• bytesmith.cabal +4/−4
• src/Data/Bytes/Parser.hs +22/−5
• src/Data/Bytes/Parser/Internal.hs +0/−2
• src/Data/Bytes/Parser/Latin.hs +46/−19
• src/Data/Bytes/Parser/Rebindable.hs +30/−23
• src/Data/Bytes/Parser/Utf8.hs +13/−8
• test/Main.hs +8/−0

@@ -1,5 +1,9 @@ # Revision history for bytesmith +## 0.3.9.0 -- 2022-07-14++* Build with GHC 9.2.3.+ ## 0.3.8.0 -- 2021-10-11  * Add `peek` and `peek'` to `Data.Bytes.Parser.Latin`.

@@ -1,6 +1,6 @@ cabal-version: 2.2 name: bytesmith-version: 0.3.8.0+version: 0.3.9.0 synopsis: Nonresumable byte parser description:   Parse bytes as fast as possible. This is a nonresumable parser@@ -33,9 +33,9 @@     Data.Bytes.Parser.Types   build-depends:     , base >=4.12 && <5-    , bytestring >=0.10.8 && <0.11-    , byteslice >=0.1.4 && <0.3-    , contiguous >= 0.4 && < 0.7+    , bytestring >=0.10.8 && <=0.11.3.0+    , byteslice >=0.2.6 && <0.3+    , contiguous >= 0.6 && < 0.7     , primitive >=0.7 && <0.8     , text-short >=0.1.3 && <0.2     , run-st >=0.1 && <0.2

@@ -15,6 +15,7 @@ {-# language TypeApplications #-} {-# language UnboxedSums #-} {-# language UnboxedTuples #-}+{-# language CPP #-}  -- | Parse non-resumable sequence of bytes. To parse a byte sequence -- as text, use the @Ascii@, @Latin@, and @Utf8@ modules instead.@@ -217,12 +218,20 @@ -- | Consume input matching the @NUL@-terminated C String. cstring :: e -> CString -> Parser e s () cstring e (Exts.Ptr ptr0) = Parser-  ( \(# arr, off0, len0 #) s -> -    let go !ptr !off !len = case Exts.indexWord8OffAddr# ptr 0# of+  ( \(# arr, off0, len0 #) s ->+    let go !ptr !off !len = case+#if MIN_VERSION_base(4,16,0)+                 Exts.word8ToWord#+#endif+            (Exts.indexWord8OffAddr# ptr 0#) of           0## -> (# s, (# | (# (), off, len #) #) #)           c -> case len of             0# -> (# s, (# e | #) #)-            _ -> case Exts.eqWord# c (Exts.indexWord8Array# arr off) of+            _ -> case Exts.eqWord# c (+#if MIN_VERSION_base(4,16,0)+                 Exts.word8ToWord#+#endif+              (Exts.indexWord8Array# arr off)) of               1# -> go (Exts.plusAddr# ptr 1# ) (off +# 1# ) (len -# 1# )               _ -> (# s, (# e | #) #)      in go ptr0 off0 len0@@ -483,7 +492,11 @@   (\x s0 -> case f x s0 of     (# s1, r #) -> case r of       (# e | #) -> (# s1, (# e | #) #)-      (# | (# W32# a, b, c #) #) -> (# s1, (# | (# a, b, c #) #) #)+      (# | (# W32# a, b, c #) #) -> (# s1, (# | (#+#if MIN_VERSION_base(4,16,0)+        Exts.word32ToWord#+#endif+        a, b, c #) #) #)   )  -- | Convert a @(Int,Int)@ parser to a @(# Int#, Int# #)@ parser.@@ -504,7 +517,11 @@   (\x s0 -> case f x s0 of     (# s1, r #) -> case r of       (# e | #) -> (# s1, (# e | #) #)-      (# | (# a, b, c #) #) -> (# s1, (# | (# W32# a, b, c #) #) #)+      (# | (# a, b, c #) #) -> (# s1, (# | (# W32# (+#if MIN_VERSION_base(4,16,0)+        Exts.wordToWord32#+#endif+        a), b, c #) #) #)   )  -- | Convert a @(# Int#, Int# #)@ parser to a @(Int,Int)@ parser.

@@ -131,9 +131,7 @@   (<*>) = Control.Monad.ap  instance Monad (Parser e s) where-  {-# inline return #-}   {-# inline (>>=) #-}-  return = pureParser   (>>=) = bindParser  instance Functor (Parser e s) where

@@ -15,6 +15,7 @@ {-# language TypeApplications #-} {-# language UnboxedSums #-} {-# language UnboxedTuples #-}+{-# language CPP #-}  -- | Parse input as though it were text encoded by -- ISO 8859-1 (Latin-1). All byte sequences are valid@@ -385,7 +386,7 @@   -> Bytes -- chunk   -> Result# e () skipDigitsAscii1LoopStart e !c = if length c > 0-  then +  then     let w = indexLatinCharArray (array c) (offset c)      in if w >= '0' && w <= '9'           then upcastUnitSuccess (skipDigitsAsciiLoop (Bytes.unsafeDrop 1 c))@@ -406,6 +407,7 @@   upcastUnitSuccess (skipDigitsAsciiLoop c)  unI :: Int -> Int#+{-# inline unI #-} unI (I# w) = w  -- | Skip the character any number of times. This succeeds@@ -600,6 +602,7 @@   else (# | (# unW acc, unI (offset chunk0), 0# #) #)  unW :: Word -> Word#+{-# inline unW #-} unW (W# w) = w  decWordStart ::@@ -619,19 +622,31 @@ upcastWord16Result :: Result# e Word# -> Result# e Word16 {-# inline upcastWord16Result #-} upcastWord16Result (# e | #) = (# e | #)-upcastWord16Result (# | (# a, b, c #) #) = (# | (# W16# a, b, c #) #)+upcastWord16Result (# | (# a, b, c #) #) = (# | (# W16# (+#if MIN_VERSION_base(4,16,0)+  Exts.wordToWord16#+#endif+  a), b, c #) #)  -- Precondition: the word is small enough upcastWord32Result :: Result# e Word# -> Result# e Word32 {-# inline upcastWord32Result #-} upcastWord32Result (# e | #) = (# e | #)-upcastWord32Result (# | (# a, b, c #) #) = (# | (# W32# a, b, c #) #)+upcastWord32Result (# | (# a, b, c #) #) = (# | (# W32# (+#if MIN_VERSION_base(4,16,0)+  Exts.wordToWord32#+#endif+  a), b, c #) #)  -- Precondition: the word is small enough upcastWord8Result :: Result# e Word# -> Result# e Word8 {-# inline upcastWord8Result #-} upcastWord8Result (# e | #) = (# e | #)-upcastWord8Result (# | (# a, b, c #) #) = (# | (# W8# a, b, c #) #)+upcastWord8Result (# | (# a, b, c #) #) = (# | (# W8# (+#if MIN_VERSION_base(4,16,0)+  Exts.wordToWord8#+#endif+  a), b, c #) #)  -- | Parse a decimal-encoded number. If the number is too large to be -- represented by a machine integer, this fails with the provided@@ -707,7 +722,7 @@     (\chunk0 s0 -> decPosIntStart e (boxBytes chunk0) s0)   '-' -> Parser -- minus sign     (\chunk0 s0 -> decNegIntStart e (boxBytes chunk0) s0)-  _ -> Parser -- no sign, there should be a digit here +  _ -> Parser -- no sign, there should be a digit here     (\chunk0 s0 ->       let !w = char2Word c - 48         in if w < 10@@ -722,7 +737,7 @@ decStandardInt# e = any e `bindFromLiftedToInt` \c -> case c of   '-' -> Parser -- minus sign     (\chunk0 s0 -> decNegIntStart e (boxBytes chunk0) s0)-  _ -> Parser -- no sign, there should be a digit here +  _ -> Parser -- no sign, there should be a digit here     (\chunk0 s0 ->       let !w = char2Word c - 48         in if w < 10@@ -759,7 +774,7 @@   '-' -> do     x <- decUnsignedInteger e     pure $! negate x-  _ -> Parser -- no sign, there should be a digit here +  _ -> Parser -- no sign, there should be a digit here     (\chunk0 s0 ->       let !w = char2Word c - 48 in       if w < 10@@ -795,11 +810,11 @@     let !w = fromIntegral @Word8 @Word           (PM.indexByteArray (array chunk0) (offset chunk0)) - 48      in if w < 10-          then +          then             case decPosIntMore e w (maxIntAsWord + 1) (Bytes.unsafeDrop 1 chunk0) of-             (# | (# x, y, z #) #) -> +             (# | (# x, y, z #) #) ->                (# s0, (# | (# (notI# x +# 1# ), y, z #) #) #)-             (# err | #) -> +             (# err | #) ->                (# s0, (# err | #) #)           else (# s0, (# e | #) #)   else (# s0, (# e | #) #)@@ -827,7 +842,7 @@ -- exceeds the upper bound. decPosIntMore ::      e -- Error message-  -> Word -- Accumulator, precondition: less than or equal to bound +  -> Word -- Accumulator, precondition: less than or equal to bound   -> Word -- Inclusive Upper Bound, either (2^63 - 1) or 2^63   -> Bytes -- Chunk   -> Result# e Int#@@ -918,7 +933,7 @@ -- | Skip all characters until the terminator is encountered -- and then consume the matching character as well. Visually, -- @skipTrailedBy \'C\'@ advances the cursor like this:--- +-- -- >  A Z B Y C X C W -- > |->->->->-| --@@ -931,7 +946,7 @@ -- | Skip all characters until the terminator is encountered. -- This does not consume the terminator. Visually, @skipUntil \'C\'@ -- advances the cursor like this:--- +-- -- >  A Z B Y C X C W -- > |->->->-| --@@ -971,7 +986,11 @@   (\x s0 -> case runParser (hexFixedWord32# e) x s0 of     (# s1, r #) -> case r of       (# err | #) -> (# s1, (# err | #) #)-      (# | (# a, b, c #) #) -> (# s1, (# | (# W32# a, b, c #) #) #)+      (# | (# a, b, c #) #) -> (# s1, (# | (# W32# (+#if MIN_VERSION_base(4,16,0)+        Exts.wordToWord32#+#endif+        a), b, c #) #) #)   )  hexFixedWord32# :: e -> Parser e s Word#@@ -1020,7 +1039,7 @@   then     let go !off !len !acc = case len of           0 -> case acc of-            W# r -> +            W# r ->               (# | (# r               ,  unI off               ,  unI (length chunk) -# 16# #) #)@@ -1042,7 +1061,11 @@   (\x s0 -> case runParser (hexFixedWord16# e) x s0 of     (# s1, r #) -> case r of       (# err | #) -> (# s1, (# err | #) #)-      (# | (# a, b, c #) #) -> (# s1, (# | (# W16# a, b, c #) #) #)+      (# | (# a, b, c #) #) -> (# s1, (# | (# W16# (+#if MIN_VERSION_base(4,16,0)+        Exts.wordToWord16#+#endif+        a), b, c #) #) #)   )  hexFixedWord16# :: e -> Parser e s Word#@@ -1074,7 +1097,11 @@   (\x s0 -> case runParser (hexFixedWord8# e) x s0 of     (# s1, r #) -> case r of       (# err | #) -> (# s1, (# err | #) #)-      (# | (# a, b, c #) #) -> (# s1, (# | (# W8# a, b, c #) #) #)+      (# | (# a, b, c #) #) -> (# s1, (# | (# W8# (+#if MIN_VERSION_base(4,16,0)+        Exts.wordToWord8#+#endif+        a), b, c #) #) #)   )  hexFixedWord8# :: e -> Parser e s Word#@@ -1173,7 +1200,7 @@ -- word is less than or equal to the upper bound positivePushBase10 :: Word -> Word -> Word -> (Bool,Word) {-# inline positivePushBase10 #-}-positivePushBase10 (W# a) (W# b) (W# upper) = +positivePushBase10 (W# a) (W# b) (W# upper) =   let !(# ca, r0 #) = Exts.timesWord2# a 10##       !r1 = Exts.plusWord# r0 b       !cb = int2Word# (gtWord# r1 upper)@@ -1183,7 +1210,7 @@  unsignedPushBase10 :: Word -> Word -> (Bool,Word) {-# inline unsignedPushBase10 #-}-unsignedPushBase10 (W# a) (W# b) = +unsignedPushBase10 (W# a) (W# b) =   let !(# ca, r0 #) = Exts.timesWord2# a 10##       !r1 = Exts.plusWord# r0 b       !cb = int2Word# (ltWord# r1 r0)

@@ -6,6 +6,7 @@ {-# language TypeInType #-} {-# language UnboxedSums #-} {-# language UnboxedTuples #-}+{-# language CPP #-}  -- | Provides levity-polymorphic variants of @>>=@, @>>@, and @pure@ -- used to assemble parsers whose result types are unlifted. This@@ -21,10 +22,16 @@   , Pure(..)   ) where -import Prelude () +import Prelude () import GHC.Exts (TYPE,RuntimeRep(..)) import Data.Bytes.Parser.Internal (Parser(..)) +#if MIN_VERSION_base(4,16,0)+import GHC.Exts (LiftedRep)+#else+type LiftedRep = 'LiftedRep+#endif+ class Bind (ra :: RuntimeRep) (rb :: RuntimeRep) where   (>>=) :: forall e s (a :: TYPE ra) (b :: TYPE rb).     Parser e s a -> (a -> Parser e s b) -> Parser e s b@@ -160,8 +167,8 @@       (# | (# _, b, c #) #) -> g (# arr, b, c #) s1   ) -bindInt2to5Parser :: forall -  (a :: TYPE ('TupleRep '[ 'IntRep, 'IntRep])) +bindInt2to5Parser :: forall+  (a :: TYPE ('TupleRep '[ 'IntRep, 'IntRep]))   (b :: TYPE ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep]))   e s.   Parser e s a -> (a -> Parser e s b) -> Parser e s b@@ -174,7 +181,7 @@         runParser (g y) (# arr, b, c #) s1   ) -sequenceInt2to5Parser :: forall +sequenceInt2to5Parser :: forall   (a :: TYPE ('TupleRep '[ 'IntRep, 'IntRep]))   (b :: TYPE ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep]))   e s.@@ -187,25 +194,25 @@       (# | (# _, b, c #) #) -> g (# arr, b, c #) s1   ) -instance Bind 'LiftedRep 'LiftedRep where+instance Bind LiftedRep LiftedRep where   {-# inline (>>=) #-}   {-# inline (>>) #-}   (>>=) = bindParser   (>>) = sequenceParser -instance Bind 'WordRep 'LiftedRep where+instance Bind 'WordRep LiftedRep where   {-# inline (>>=) #-}   {-# inline (>>) #-}   (>>=) = bindWordParser   (>>) = sequenceWordParser -instance Bind 'IntRep 'LiftedRep where+instance Bind 'IntRep LiftedRep where   {-# inline (>>=) #-}   {-# inline (>>) #-}   (>>=) = bindIntParser   (>>) = sequenceIntParser -instance Bind ('TupleRep '[ 'IntRep, 'IntRep]) 'LiftedRep where+instance Bind ('TupleRep '[ 'IntRep, 'IntRep]) LiftedRep where   {-# inline (>>=) #-}   {-# inline (>>) #-}   (>>=) = bindIntPairParser@@ -213,15 +220,15 @@   instance Bind ('TupleRep '[ 'IntRep, 'IntRep])-              ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep]) +              ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep])   where   {-# inline (>>=) #-}   {-# inline (>>) #-}   (>>=) = bindInt2to5Parser   (>>) = sequenceInt2to5Parser -instance Bind ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep]) -              'LiftedRep+instance Bind ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep])+              LiftedRep   where   {-# inline (>>=) #-}   {-# inline (>>) #-}@@ -230,21 +237,21 @@   instance Bind 'IntRep-              ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep]) +              ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep])   where   {-# inline (>>=) #-}   {-# inline (>>) #-}   (>>=) = bindFromIntToInt5   (>>) = sequenceIntToInt5 -instance Bind 'LiftedRep ('TupleRep '[ 'IntRep, 'IntRep]) where+instance Bind LiftedRep ('TupleRep '[ 'IntRep, 'IntRep]) where   {-# inline (>>=) #-}   {-# inline (>>) #-}   (>>=) = bindFromLiftedToIntPair   (>>) = sequenceLiftedToIntPair -instance Bind 'LiftedRep-              ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep]) +instance Bind LiftedRep+              ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep])   where   {-# inline (>>=) #-}   {-# inline (>>) #-}@@ -257,13 +264,13 @@   (>>=) = bindFromIntToIntPair   (>>) = sequenceIntToIntPair -instance Bind 'LiftedRep 'IntRep where+instance Bind LiftedRep 'IntRep where   {-# inline (>>=) #-}   {-# inline (>>) #-}   (>>=) = bindFromLiftedToInt   (>>) = sequenceLiftedToInt -instance Pure 'LiftedRep where+instance Pure LiftedRep where   {-# inline pure #-}   pure = pureParser @@ -343,7 +350,7 @@  bindFromLiftedToIntPair ::      forall s e-       (a :: TYPE 'LiftedRep)+       (a :: TYPE LiftedRep)        (b :: TYPE ('TupleRep '[ 'IntRep, 'IntRep ])).      Parser s e a   -> (a -> Parser s e b)@@ -359,7 +366,7 @@  sequenceLiftedToIntPair ::      forall s e-       (a :: TYPE 'LiftedRep)+       (a :: TYPE LiftedRep)        (b :: TYPE ('TupleRep '[ 'IntRep, 'IntRep ])).      Parser s e a   -> Parser s e b@@ -375,7 +382,7 @@  bindFromLiftedToInt5 ::      forall s e-       (a :: TYPE 'LiftedRep)+       (a :: TYPE LiftedRep)        (b :: TYPE ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep])).      Parser s e a   -> (a -> Parser s e b)@@ -391,7 +398,7 @@  sequenceLiftedToInt5 ::      forall s e-       (a :: TYPE 'LiftedRep)+       (a :: TYPE LiftedRep)        (b :: TYPE ('TupleRep '[ 'IntRep, 'IntRep, 'IntRep, 'IntRep, 'IntRep ])).      Parser s e a   -> Parser s e b@@ -406,7 +413,7 @@  bindFromLiftedToInt ::      forall s e-       (a :: TYPE 'LiftedRep)+       (a :: TYPE LiftedRep)        (b :: TYPE 'IntRep).      Parser s e a   -> (a -> Parser s e b)@@ -422,7 +429,7 @@  sequenceLiftedToInt ::      forall s e-       (a :: TYPE 'LiftedRep)+       (a :: TYPE LiftedRep)        (b :: TYPE 'IntRep).      Parser s e a   -> Parser s e b

@@ -15,6 +15,7 @@ {-# language TypeApplications #-} {-# language UnboxedSums #-} {-# language UnboxedTuples #-}+{-# language CPP #-}  -- | Parse input as UTF-8-encoded text. Parsers in this module will -- fail if they encounter a byte above @0x7F@.@@ -46,8 +47,12 @@   (\(# arr, off, len #) s0 -> case len ># 0# of     1# ->       let !w0 = Exts.indexWord8Array# arr off-       in if | oneByteChar (W8# w0) -> -                 (# s0, (# | (# chr# (Exts.word2Int# w0), off +# 1#, len -# 1# #) #) #)+       in if | oneByteChar (W8# w0) ->+                 (# s0, (# | (# chr# (Exts.word2Int# (+#if MIN_VERSION_base(4,16,0)+                 Exts.word8ToWord#+#endif+                 w0)), off +# 1#, len -# 1# #) #) #)              | twoByteChar (W8# w0) ->                  if | I# len > 1                     , w1 <- Exts.indexWord8Array# arr (off +# 1#)@@ -81,9 +86,9 @@ codepointFromFourBytes w1 w2 w3 w4 = C#   ( chr#     ( unI $ fromIntegral-      ( unsafeShiftL (word8ToWord w1 .&. 0b00001111) 18 .|. -        unsafeShiftL (word8ToWord w2 .&. 0b00111111) 12 .|. -        unsafeShiftL (word8ToWord w3 .&. 0b00111111) 6 .|. +      ( unsafeShiftL (word8ToWord w1 .&. 0b00001111) 18 .|.+        unsafeShiftL (word8ToWord w2 .&. 0b00111111) 12 .|.+        unsafeShiftL (word8ToWord w3 .&. 0b00111111) 6 .|.         (word8ToWord w4 .&. 0b00111111)       )     )@@ -93,8 +98,8 @@ codepointFromThreeBytes w1 w2 w3 = C#   ( chr#     ( unI $ fromIntegral-      ( unsafeShiftL (word8ToWord w1 .&. 0b00001111) 12 .|. -        unsafeShiftL (word8ToWord w2 .&. 0b00111111) 6 .|. +      ( unsafeShiftL (word8ToWord w1 .&. 0b00001111) 12 .|.+        unsafeShiftL (word8ToWord w2 .&. 0b00111111) 6 .|.         (word8ToWord w3 .&. 0b00111111)       )     )@@ -104,7 +109,7 @@ codepointFromTwoBytes w1 w2 = C#   ( chr#     ( unI $ fromIntegral @Word @Int-      ( unsafeShiftL (word8ToWord w1 .&. 0b00011111) 6 .|. +      ( unsafeShiftL (word8ToWord w1 .&. 0b00011111) 6 .|.         (word8ToWord w2 .&. 0b00111111)       )     )

@@ -1,5 +1,6 @@ {-# language BangPatterns #-} {-# language DataKinds #-}+{-# language MagicHash #-} {-# language MultiWayIf #-} {-# language NumDecimals #-} {-# language OverloadedStrings #-}@@ -111,6 +112,13 @@       P.parseBytes (replicateM (length xs) (BigEndian.word128 ())) bs       ===       P.parseBytes (fmap Exts.toList (BigEndian.word128Array () (length xs))) bs+  , testProperty "cstring" $ \(xs :: [Word8]) ->+      let ys = Exts.fromList xs+          bs = Bytes.singleton 0x31 <> ys+       in+      P.parseBytes (P.cstring () (Exts.Ptr "1"# ) *> P.bytes () ys *> pure 42) bs+      ===+      (P.Success (Slice (Bytes.length ys + 1) 0 42) :: P.Result () Integer)   , testCase "big-endian-word256" $       P.parseBytesMaybe (BigEndian.word256Array () 1) (Exts.fromList [         0x12, 0x34, 0x56, 0x78, 0x90,