1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
|
%
% (c) The University of Glasgow 2006
% (c) The University of Glasgow, 1997-2006
%
Buffers for scanning string input stored in external arrays.
\begin{code}
module StringBuffer
(
StringBuffer(..),
-- non-abstract for vs\/HaskellService
-- * Creation\/destruction
hGetStringBuffer,
hGetStringBufferBlock,
appendStringBuffers,
stringToStringBuffer,
-- * Inspection
nextChar,
currentChar,
prevChar,
atEnd,
-- * Moving and comparison
stepOn,
offsetBytes,
byteDiff,
-- * Conversion
lexemeToString,
lexemeToFastString,
-- * Parsing integers
parseUnsignedInteger,
) where
#include "HsVersions.h"
import Encoding
import FastString ( FastString,mkFastString,mkFastStringBytes )
import Foreign
import System.IO ( hGetBuf, hFileSize,IOMode(ReadMode), hClose
, Handle, hTell )
import GHC.Exts
import GHC.IOBase ( IO(..) )
import GHC.Base ( unsafeChr )
#if __GLASGOW_HASKELL__ >= 601
import System.IO ( openBinaryFile )
#else
import IOExts ( openFileEx, IOModeEx(..) )
#endif
#if __GLASGOW_HASKELL__ < 601
openBinaryFile fp mode = openFileEx fp (BinaryMode mode)
#endif
-- -----------------------------------------------------------------------------
-- The StringBuffer type
-- |A StringBuffer is an internal pointer to a sized chunk of bytes.
-- The bytes are intended to be *immutable*. There are pure
-- operations to read the contents of a StringBuffer.
--
-- A StringBuffer may have a finalizer, depending on how it was
-- obtained.
--
data StringBuffer
= StringBuffer {
buf :: {-# UNPACK #-} !(ForeignPtr Word8),
len :: {-# UNPACK #-} !Int, -- length
cur :: {-# UNPACK #-} !Int -- current pos
}
-- The buffer is assumed to be UTF-8 encoded, and furthermore
-- we add three '\0' bytes to the end as sentinels so that the
-- decoder doesn't have to check for overflow at every single byte
-- of a multibyte sequence.
instance Show StringBuffer where
showsPrec _ s = showString "<stringbuffer("
. shows (len s) . showString "," . shows (cur s)
. showString ">"
-- -----------------------------------------------------------------------------
-- Creation / Destruction
hGetStringBuffer :: FilePath -> IO StringBuffer
hGetStringBuffer fname = do
h <- openBinaryFile fname ReadMode
size_i <- hFileSize h
let size = fromIntegral size_i
buf <- mallocForeignPtrArray (size+3)
withForeignPtr buf $ \ptr -> do
r <- if size == 0 then return 0 else hGetBuf h ptr size
hClose h
if (r /= size)
then ioError (userError "short read of file")
else do
pokeArray (ptr `plusPtr` size :: Ptr Word8) [0,0,0]
-- sentinels for UTF-8 decoding
return (StringBuffer buf size 0)
hGetStringBufferBlock :: Handle -> Int -> IO StringBuffer
hGetStringBufferBlock handle wanted
= do size_i <- hFileSize handle
offset_i <- hTell handle
let size = min wanted (fromIntegral $ size_i-offset_i)
buf <- mallocForeignPtrArray (size+3)
withForeignPtr buf $ \ptr ->
do r <- if size == 0 then return 0 else hGetBuf handle ptr size
if r /= size
then ioError (userError $ "short read of file: "++show(r,size,size_i,handle))
else do pokeArray (ptr `plusPtr` size :: Ptr Word8) [0,0,0]
return (StringBuffer buf size 0)
appendStringBuffers :: StringBuffer -> StringBuffer -> IO StringBuffer
appendStringBuffers sb1 sb2
= do newBuf <- mallocForeignPtrArray (size+3)
withForeignPtr newBuf $ \ptr ->
withForeignPtr (buf sb1) $ \sb1Ptr ->
withForeignPtr (buf sb2) $ \sb2Ptr ->
do copyArray (sb1Ptr `advancePtr` cur sb1) ptr (calcLen sb1)
copyArray (sb2Ptr `advancePtr` cur sb2) (ptr `advancePtr` cur sb1) (calcLen sb2)
pokeArray (ptr `advancePtr` size) [0,0,0]
return (StringBuffer newBuf size 0)
where calcLen sb = len sb - cur sb
size = calcLen sb1 + calcLen sb2
stringToStringBuffer :: String -> IO StringBuffer
stringToStringBuffer str = do
let size = utf8EncodedLength str
buf <- mallocForeignPtrArray (size+3)
withForeignPtr buf $ \ptr -> do
utf8EncodeString ptr str
pokeArray (ptr `plusPtr` size :: Ptr Word8) [0,0,0]
-- sentinels for UTF-8 decoding
return (StringBuffer buf size 0)
-- -----------------------------------------------------------------------------
-- Grab a character
-- Getting our fingers dirty a little here, but this is performance-critical
{-# INLINE nextChar #-}
nextChar :: StringBuffer -> (Char,StringBuffer)
nextChar (StringBuffer buf len (I# cur#)) =
inlinePerformIO $ do
withForeignPtr buf $ \(Ptr a#) -> do
case utf8DecodeChar# (a# `plusAddr#` cur#) of
(# c#, b# #) ->
let cur' = I# (b# `minusAddr#` a#) in
return (C# c#, StringBuffer buf len cur')
currentChar :: StringBuffer -> Char
currentChar = fst . nextChar
prevChar :: StringBuffer -> Char -> Char
prevChar (StringBuffer buf len 0) deflt = deflt
prevChar (StringBuffer buf len cur) deflt =
inlinePerformIO $ do
withForeignPtr buf $ \p -> do
p' <- utf8PrevChar (p `plusPtr` cur)
return (fst (utf8DecodeChar p'))
-- -----------------------------------------------------------------------------
-- Moving
stepOn :: StringBuffer -> StringBuffer
stepOn s = snd (nextChar s)
offsetBytes :: Int -> StringBuffer -> StringBuffer
offsetBytes i s = s { cur = cur s + i }
byteDiff :: StringBuffer -> StringBuffer -> Int
byteDiff s1 s2 = cur s2 - cur s1
atEnd :: StringBuffer -> Bool
atEnd (StringBuffer _ l c) = l == c
-- -----------------------------------------------------------------------------
-- Conversion
lexemeToString :: StringBuffer -> Int {-bytes-} -> String
lexemeToString _ 0 = ""
lexemeToString (StringBuffer buf _ cur) bytes =
inlinePerformIO $
withForeignPtr buf $ \ptr ->
utf8DecodeString (ptr `plusPtr` cur) bytes
lexemeToFastString :: StringBuffer -> Int {-bytes-} -> FastString
lexemeToFastString _ 0 = mkFastString ""
lexemeToFastString (StringBuffer buf _ cur) len =
inlinePerformIO $
withForeignPtr buf $ \ptr ->
return $! mkFastStringBytes (ptr `plusPtr` cur) len
-- -----------------------------------------------------------------------------
-- Parsing integer strings in various bases
byteOff :: StringBuffer -> Int -> Char
byteOff (StringBuffer buf _ cur) i =
inlinePerformIO $ withForeignPtr buf $ \ptr -> do
w <- peek (ptr `plusPtr` (cur+i))
return (unsafeChr (fromIntegral (w::Word8)))
-- | XXX assumes ASCII digits only (by using byteOff)
parseUnsignedInteger :: StringBuffer -> Int -> Integer -> (Char->Int) -> Integer
parseUnsignedInteger buf len radix char_to_int
= go 0 0
where
go i x | i == len = x
| otherwise = go (i+1)
(x * radix + toInteger (char_to_int (byteOff buf i)))
-- -----------------------------------------------------------------------------
-- under the carpet
-- Just like unsafePerformIO, but we inline it.
{-# INLINE inlinePerformIO #-}
inlinePerformIO :: IO a -> a
inlinePerformIO (IO m) = case m realWorld# of (# _, r #) -> r
\end{code}
|
