FastString.lhs 17.4 KB
Newer Older
1
%
2
% (c) The University of Glasgow, 1997-2006
3
%
4 5
\begin{code}
{-
Ian Lynagh's avatar
Ian Lynagh committed
6 7 8 9
FastString:     A compact, hash-consed, representation of character strings.
                Comparison is O(1), and you can get a Unique from them.
                Generated by the FSLIT macro
                Turn into SDoc with Outputable.ftext
10

Ian Lynagh's avatar
Ian Lynagh committed
11 12 13 14 15
LitString:      Just a wrapper for the Addr# of a C string (Ptr CChar).
                Practically no operations
                Outputing them is fast
                Generated by the SLIT macro
                Turn into SDoc with Outputable.ptext
16 17

Use LitString unless you want the facilities of FastString
18
-}
19 20
module FastString
       (
Ian Lynagh's avatar
Ian Lynagh committed
21 22
        -- * FastStrings
        FastString(..),     -- not abstract, for now.
23

Ian Lynagh's avatar
Ian Lynagh committed
24
        -- ** Construction
25
        mkFastString,
Ian Lynagh's avatar
Ian Lynagh committed
26
        mkFastStringBytes,
27
        mkFastStringByteList,
Ian Lynagh's avatar
Ian Lynagh committed
28
        mkFastStringForeignPtr,
29
#if defined(__GLASGOW_HASKELL__)
Ian Lynagh's avatar
Ian Lynagh committed
30
        mkFastString#,
31
#endif
Ian Lynagh's avatar
Ian Lynagh committed
32 33
        mkZFastString,
        mkZFastStringBytes,
34

Ian Lynagh's avatar
Ian Lynagh committed
35 36 37
        -- ** Deconstruction
        unpackFS,           -- :: FastString -> String
        bytesFS,            -- :: FastString -> [Word8]
38

Ian Lynagh's avatar
Ian Lynagh committed
39 40 41
        -- ** Encoding
        isZEncoded,
        zEncodeFS,
42

Ian Lynagh's avatar
Ian Lynagh committed
43
        -- ** Operations
44
        uniqueOfFS,
Ian Lynagh's avatar
Ian Lynagh committed
45 46 47
        lengthFS,
        nullFS,
        appendFS,
48 49
        headFS,
        tailFS,
Ian Lynagh's avatar
Ian Lynagh committed
50
        concatFS,
51
        consFS,
Ian Lynagh's avatar
Ian Lynagh committed
52
        nilFS,
53

Ian Lynagh's avatar
Ian Lynagh committed
54
        -- ** Outputing
55 56
        hPutFS,

Ian Lynagh's avatar
Ian Lynagh committed
57 58 59
        -- ** Internal
        getFastStringTable,
        hasZEncoding,
60

Ian Lynagh's avatar
Ian Lynagh committed
61 62
        -- * LitStrings
        LitString,
63
#if defined(__GLASGOW_HASKELL__)
Ian Lynagh's avatar
Ian Lynagh committed
64
        mkLitString#,
65
#else
Ian Lynagh's avatar
Ian Lynagh committed
66
        mkLitString,
67
#endif
Ian Lynagh's avatar
Ian Lynagh committed
68 69
        unpackLitString,
        strLength,
70

Ian Lynagh's avatar
Ian Lynagh committed
71
        ptrStrLength
72 73
       ) where

74
#include "HsVersions.h"
75

76
import Encoding
77 78
import FastTypes
import FastFunctions
79
import Panic
80

81 82
import Foreign
import Foreign.C
83
import GHC.Exts
84
import System.IO
85
import System.IO.Unsafe ( unsafePerformIO )
Ian Lynagh's avatar
Ian Lynagh committed
86 87
import Data.IORef       ( IORef, newIORef, readIORef, writeIORef )
import Data.Maybe       ( isJust )
88
#if !defined(__GLASGOW_HASKELL__)
Ian Lynagh's avatar
Ian Lynagh committed
89
import Data.Char        ( ord )
90
#endif
91

Ian Lynagh's avatar
Ian Lynagh committed
92 93
import GHC.IOBase       ( IO(..) )
import GHC.Ptr          ( Ptr(..) )
94

95 96
#define hASH_TBL_SIZE          4091
#define hASH_TBL_SIZE_UNBOXED  4091#
97 98


99 100 101 102 103
{-|
A 'FastString' is an array of bytes, hashed to support fast O(1)
comparison.  It is also associated with a character encoding, so that
we know how to convert a 'FastString' to the local encoding, or to the
Z-encoding used by the compiler internally.
104

105 106 107 108
'FastString's support a memoized conversion to the Z-encoding via zEncodeFS.
-}

data FastString = FastString {
Ian Lynagh's avatar
Ian Lynagh committed
109 110 111
      uniq    :: {-# UNPACK #-} !Int, -- unique id
      n_bytes :: {-# UNPACK #-} !Int, -- number of bytes
      n_chars :: {-# UNPACK #-} !Int, -- number of chars
112 113 114 115 116
      buf     :: {-# UNPACK #-} !(ForeignPtr Word8),
      enc     :: FSEncoding
  }

data FSEncoding
Ian Lynagh's avatar
Ian Lynagh committed
117
    -- including strings that don't need any encoding
118
  = ZEncoded
Ian Lynagh's avatar
Ian Lynagh committed
119
    -- A UTF-8 string with a memoized Z-encoding
120
  | UTF8Encoded {-# UNPACK #-} !(IORef (Maybe FastString))
121

122 123
instance Eq FastString where
  f1 == f2  =  uniq f1 == uniq f2
124

125
instance Ord FastString where
Ian Lynagh's avatar
Ian Lynagh committed
126
    -- Compares lexicographically, not by unique
127
    a <= b = case cmpFS a b of { LT -> True;  EQ -> True;  GT -> False }
Ian Lynagh's avatar
Ian Lynagh committed
128
    a <  b = case cmpFS a b of { LT -> True;  EQ -> False; GT -> False }
129
    a >= b = case cmpFS a b of { LT -> False; EQ -> True;  GT -> True  }
Ian Lynagh's avatar
Ian Lynagh committed
130 131 132 133 134
    a >  b = case cmpFS a b of { LT -> False; EQ -> False; GT -> True  }
    max x y | x >= y    =  x
            | otherwise =  y
    min x y | x <= y    =  x
            | otherwise =  y
135
    compare a b = cmpFS a b
136

137 138 139
instance Show FastString where
   show fs = show (unpackFS fs)

140 141 142
cmpFS :: FastString -> FastString -> Ordering
cmpFS (FastString u1 l1 _ buf1 _) (FastString u2 l2 _ buf2 _) =
  if u1 == u2 then EQ else
Isaac Dupree's avatar
Isaac Dupree committed
143 144 145 146 147 148 149 150 151 152 153
  case unsafeMemcmp buf1 buf2 (min l1 l2) `compare` 0 of
     LT -> LT
     EQ -> compare l1 l2
     GT -> GT

unsafeMemcmp :: ForeignPtr a -> ForeignPtr b -> Int -> Int
unsafeMemcmp buf1 buf2 l =
      inlinePerformIO $
        withForeignPtr buf1 $ \p1 ->
        withForeignPtr buf2 $ \p2 ->
          memcmp p1 p2 l
154

155
#ifndef __HADDOCK__
Ian Lynagh's avatar
Ian Lynagh committed
156
foreign import ccall unsafe "ghc_memcmp"
157 158
  memcmp :: Ptr a -> Ptr b -> Int -> IO Int
#endif
159

160 161
-- -----------------------------------------------------------------------------
-- Construction
162

163
{-
164 165 166 167
Internally, the compiler will maintain a fast string symbol
table, providing sharing and fast comparison. Creation of
new @FastString@s then covertly does a lookup, re-using the
@FastString@ if there was a hit.
168
-}
169

Ian Lynagh's avatar
Ian Lynagh committed
170
data FastStringTable =
171
 FastStringTable
172
    {-# UNPACK #-} !Int
173
    (MutableArray# RealWorld [FastString])
174

175
{-# NOINLINE string_table #-}
176
string_table :: IORef FastStringTable
Ian Lynagh's avatar
Ian Lynagh committed
177
string_table =
178
 unsafePerformIO $ do
179 180 181 182
   tab <- IO $ \s1# -> case newArray# hASH_TBL_SIZE_UNBOXED [] s1# of
                           (# s2#, arr# #) ->
                               (# s2#, FastStringTable 0 arr# #)
   newIORef tab
183 184 185 186 187 188 189 190 191

lookupTbl :: FastStringTable -> Int -> IO [FastString]
lookupTbl (FastStringTable _ arr#) (I# i#) =
  IO $ \ s# -> readArray# arr# i# s#

updTbl :: IORef FastStringTable -> FastStringTable -> Int -> [FastString] -> IO ()
updTbl fs_table_var (FastStringTable uid arr#) (I# i#) ls = do
  (IO $ \ s# -> case writeArray# arr# i# ls s# of { s2# -> (# s2#, () #) })
  writeIORef fs_table_var (FastStringTable (uid+1) arr#)
192

193
mkFastString# :: Addr# -> FastString
194
mkFastString# a# = mkFastStringBytes ptr (ptrStrLength ptr)
195
  where ptr = Ptr a#
196

197 198
mkFastStringBytes :: Ptr Word8 -> Int -> FastString
mkFastStringBytes ptr len = unsafePerformIO $ do
199
  ft@(FastStringTable uid _) <- readIORef string_table
200
  let
201 202
   h = hashStr ptr len
   add_it ls = do
Ian Lynagh's avatar
Ian Lynagh committed
203 204 205 206
        fs <- copyNewFastString uid ptr len
        updTbl string_table ft h (fs:ls)
        {- _trace ("new: " ++ show f_str)   $ -}
        return fs
207 208
  --
  lookup_result <- lookupTbl ft h
sof's avatar
sof committed
209
  case lookup_result of
210 211 212 213
    [] -> add_it []
    ls -> do
       b <- bucket_match ls len ptr
       case b of
Ian Lynagh's avatar
Ian Lynagh committed
214 215
         Nothing -> add_it ls
         Just v  -> {- _trace ("re-use: "++show v) $ -} return v
216 217 218

mkZFastStringBytes :: Ptr Word8 -> Int -> FastString
mkZFastStringBytes ptr len = unsafePerformIO $ do
219
  ft@(FastStringTable uid _) <- readIORef string_table
sof's avatar
sof committed
220
  let
221 222
   h = hashStr ptr len
   add_it ls = do
Ian Lynagh's avatar
Ian Lynagh committed
223 224 225 226
        fs <- copyNewZFastString uid ptr len
        updTbl string_table ft h (fs:ls)
        {- _trace ("new: " ++ show f_str)   $ -}
        return fs
227 228
  --
  lookup_result <- lookupTbl ft h
sof's avatar
sof committed
229
  case lookup_result of
230 231 232 233
    [] -> add_it []
    ls -> do
       b <- bucket_match ls len ptr
       case b of
Ian Lynagh's avatar
Ian Lynagh committed
234 235
         Nothing -> add_it ls
         Just v  -> {- _trace ("re-use: "++show v) $ -} return v
236 237 238 239 240 241

-- | Create a 'FastString' from an existing 'ForeignPtr'; the difference
-- between this and 'mkFastStringBytes' is that we don't have to copy
-- the bytes if the string is new to the table.
mkFastStringForeignPtr :: Ptr Word8 -> ForeignPtr Word8 -> Int -> IO FastString
mkFastStringForeignPtr ptr fp len = do
242
  ft@(FastStringTable uid _) <- readIORef string_table
243
--  _trace ("hashed: "++show (I# h)) $
244
  let
245 246
    h = hashStr ptr len
    add_it ls = do
Ian Lynagh's avatar
Ian Lynagh committed
247 248 249 250
        fs <- mkNewFastString uid ptr fp len
        updTbl string_table ft h (fs:ls)
        {- _trace ("new: " ++ show f_str)   $ -}
        return fs
251 252 253 254 255 256 257
  --
  lookup_result <- lookupTbl ft h
  case lookup_result of
    [] -> add_it []
    ls -> do
       b <- bucket_match ls len ptr
       case b of
Ian Lynagh's avatar
Ian Lynagh committed
258 259
         Nothing -> add_it ls
         Just v  -> {- _trace ("re-use: "++show v) $ -} return v
260 261 262

mkZFastStringForeignPtr :: Ptr Word8 -> ForeignPtr Word8 -> Int -> IO FastString
mkZFastStringForeignPtr ptr fp len = do
263
  ft@(FastStringTable uid _) <- readIORef string_table
264 265 266 267
--  _trace ("hashed: "++show (I# h)) $
  let
    h = hashStr ptr len
    add_it ls = do
Ian Lynagh's avatar
Ian Lynagh committed
268 269 270 271
        fs <- mkNewZFastString uid ptr fp len
        updTbl string_table ft h (fs:ls)
        {- _trace ("new: " ++ show f_str)   $ -}
        return fs
272 273
  --
  lookup_result <- lookupTbl ft h
274
  case lookup_result of
275 276 277 278
    [] -> add_it []
    ls -> do
       b <- bucket_match ls len ptr
       case b of
Ian Lynagh's avatar
Ian Lynagh committed
279 280
         Nothing -> add_it ls
         Just v  -> {- _trace ("re-use: "++show v) $ -} return v
281 282


283 284
-- | Creates a UTF-8 encoded 'FastString' from a 'String'
mkFastString :: String -> FastString
Ian Lynagh's avatar
Ian Lynagh committed
285
mkFastString str =
286 287 288 289 290
  inlinePerformIO $ do
    let l = utf8EncodedLength str
    buf <- mallocForeignPtrBytes l
    withForeignPtr buf $ \ptr -> do
      utf8EncodeString ptr str
Ian Lynagh's avatar
Ian Lynagh committed
291
      mkFastStringForeignPtr ptr buf l
292

293 294
-- | Creates a 'FastString' from a UTF-8 encoded @[Word8]@
mkFastStringByteList :: [Word8] -> FastString
Ian Lynagh's avatar
Ian Lynagh committed
295
mkFastStringByteList str =
296 297 298 299 300
  inlinePerformIO $ do
    let l = Prelude.length str
    buf <- mallocForeignPtrBytes l
    withForeignPtr buf $ \ptr -> do
      pokeArray (castPtr ptr) str
Ian Lynagh's avatar
Ian Lynagh committed
301
      mkFastStringForeignPtr ptr buf l
302 303 304

-- | Creates a Z-encoded 'FastString' from a 'String'
mkZFastString :: String -> FastString
Ian Lynagh's avatar
Ian Lynagh committed
305
mkZFastString str =
306 307 308 309 310
  inlinePerformIO $ do
    let l = Prelude.length str
    buf <- mallocForeignPtrBytes l
    withForeignPtr buf $ \ptr -> do
      pokeCAString (castPtr ptr) str
Ian Lynagh's avatar
Ian Lynagh committed
311
      mkZFastStringForeignPtr ptr buf l
312

313
bucket_match :: [FastString] -> Int -> Ptr Word8 -> IO (Maybe FastString)
314 315 316
bucket_match [] _ _ = return Nothing
bucket_match (v@(FastString _ l _ buf _):ls) len ptr
      | len == l  =  do
Ian Lynagh's avatar
Ian Lynagh committed
317 318 319 320 321
         b <- cmpStringPrefix ptr buf len
         if b then return (Just v)
              else bucket_match ls len ptr
      | otherwise =
         bucket_match ls len ptr
322

323 324
mkNewFastString :: Int -> Ptr Word8 -> ForeignPtr Word8 -> Int
                -> IO FastString
325 326 327 328 329
mkNewFastString uid ptr fp len = do
  ref <- newIORef Nothing
  n_chars <- countUTF8Chars ptr len
  return (FastString uid len n_chars fp (UTF8Encoded ref))

330 331 332
mkNewZFastString :: Int -> Ptr Word8 -> ForeignPtr Word8 -> Int
                 -> IO FastString
mkNewZFastString uid _ fp len = do
333 334
  return (FastString uid len len fp ZEncoded)

335
copyNewFastString :: Int -> Ptr Word8 -> Int -> IO FastString
336 337 338 339 340 341
copyNewFastString uid ptr len = do
  fp <- copyBytesToForeignPtr ptr len
  ref <- newIORef Nothing
  n_chars <- countUTF8Chars ptr len
  return (FastString uid len n_chars fp (UTF8Encoded ref))

342
copyNewZFastString :: Int -> Ptr Word8 -> Int -> IO FastString
343 344 345 346
copyNewZFastString uid ptr len = do
  fp <- copyBytesToForeignPtr ptr len
  return (FastString uid len len fp ZEncoded)

347
copyBytesToForeignPtr :: Ptr Word8 -> Int -> IO (ForeignPtr Word8)
348 349 350 351 352 353 354 355 356 357 358 359 360
copyBytesToForeignPtr ptr len = do
  fp <- mallocForeignPtrBytes len
  withForeignPtr fp $ \ptr' -> copyBytes ptr' ptr len
  return fp

cmpStringPrefix :: Ptr Word8 -> ForeignPtr Word8 -> Int -> IO Bool
cmpStringPrefix ptr fp len =
  withForeignPtr fp $ \ptr' -> do
    r <- memcmp ptr ptr' len
    return (r == 0)


hashStr  :: Ptr Word8 -> Int -> Int
361
 -- use the Addr to produce a hash value between 0 & m (inclusive)
362
hashStr (Ptr a#) (I# len#) = loop 0# 0#
Ian Lynagh's avatar
Ian Lynagh committed
363
   where
364
    loop h n | n GHC.Exts.==# len# = I# h
Ian Lynagh's avatar
Ian Lynagh committed
365 366 367 368
             | otherwise  = loop h2 (n GHC.Exts.+# 1#)
          where c = ord# (indexCharOffAddr# a# n)
                h2 = (c GHC.Exts.+# (h GHC.Exts.*# 128#)) `remInt#`
                     hASH_TBL_SIZE#
369

370 371
-- -----------------------------------------------------------------------------
-- Operations
372

373 374 375
-- | Returns the length of the 'FastString' in characters
lengthFS :: FastString -> Int
lengthFS f = n_chars f
376

377 378 379
-- | Returns 'True' if the 'FastString' is Z-encoded
isZEncoded :: FastString -> Bool
isZEncoded fs | ZEncoded <- enc fs = True
Ian Lynagh's avatar
Ian Lynagh committed
380
              | otherwise          = False
381

382 383 384
-- | Returns 'True' if this 'FastString' is not Z-encoded but already has
-- a Z-encoding cached (used in producing stats).
hasZEncoding :: FastString -> Bool
385
hasZEncoding (FastString _ _ _ _ enc) =
386 387 388 389 390
  case enc of
    ZEncoded -> False
    UTF8Encoded ref ->
      inlinePerformIO $ do
        m <- readIORef ref
Ian Lynagh's avatar
Ian Lynagh committed
391
        return (isJust m)
392

393 394 395 396 397 398
-- | Returns 'True' if the 'FastString' is empty
nullFS :: FastString -> Bool
nullFS f  =  n_bytes f == 0

-- | unpacks and decodes the FastString
unpackFS :: FastString -> String
Ian Lynagh's avatar
Ian Lynagh committed
399
unpackFS (FastString _ n_bytes _ buf enc) =
400 401
  inlinePerformIO $ withForeignPtr buf $ \ptr ->
    case enc of
Ian Lynagh's avatar
Ian Lynagh committed
402 403
        ZEncoded      -> peekCAStringLen (castPtr ptr,n_bytes)
        UTF8Encoded _ -> utf8DecodeString ptr n_bytes
404 405

bytesFS :: FastString -> [Word8]
406
bytesFS (FastString _ n_bytes _ buf _) =
407 408 409 410 411 412 413 414 415
  inlinePerformIO $ withForeignPtr buf $ \ptr ->
    peekArray n_bytes ptr

-- | returns a Z-encoded version of a 'FastString'.  This might be the
-- original, if it was already Z-encoded.  The first time this
-- function is applied to a particular 'FastString', the results are
-- memoized.
--
zEncodeFS :: FastString -> FastString
416
zEncodeFS fs@(FastString _ _ _ _ enc) =
417 418 419 420 421 422
  case enc of
    ZEncoded -> fs
    UTF8Encoded ref ->
      inlinePerformIO $ do
        m <- readIORef ref
        case m of
Ian Lynagh's avatar
Ian Lynagh committed
423 424
          Just fs -> return fs
          Nothing -> do
425
            let efs = mkZFastString (zEncodeString (unpackFS fs))
Ian Lynagh's avatar
Ian Lynagh committed
426 427
            writeIORef ref (Just efs)
            return efs
428 429 430 431 432 433 434 435

appendFS :: FastString -> FastString -> FastString
appendFS fs1 fs2 = mkFastString (unpackFS fs1 ++ unpackFS fs2)

concatFS :: [FastString] -> FastString
concatFS ls = mkFastString (Prelude.concat (map unpackFS ls)) -- ToDo: do better

headFS :: FastString -> Char
436 437
headFS (FastString _ 0 _ _ _) = panic "headFS: Empty FastString"
headFS (FastString _ _ _ buf enc) =
438 439
  inlinePerformIO $ withForeignPtr buf $ \ptr -> do
    case enc of
Ian Lynagh's avatar
Ian Lynagh committed
440 441 442 443 444
      ZEncoded -> do
         w <- peek (castPtr ptr)
         return (castCCharToChar w)
      UTF8Encoded _ ->
         return (fst (utf8DecodeChar ptr))
445 446

tailFS :: FastString -> FastString
447
tailFS (FastString _ 0 _ _ _) = panic "tailFS: Empty FastString"
Ian Lynagh's avatar
Ian Lynagh committed
448
tailFS (FastString _ n_bytes _ buf enc) =
449 450 451
  inlinePerformIO $ withForeignPtr buf $ \ptr -> do
    case enc of
      ZEncoded -> do
Ian Lynagh's avatar
Ian Lynagh committed
452
        return $! mkZFastStringBytes (ptr `plusPtr` 1) (n_bytes - 1)
453
      UTF8Encoded _ -> do
Ian Lynagh's avatar
Ian Lynagh committed
454 455 456
         let (_,ptr') = utf8DecodeChar ptr
         let off = ptr' `minusPtr` ptr
         return $! mkFastStringBytes (ptr `plusPtr` off) (n_bytes - off)
457 458 459 460

consFS :: Char -> FastString -> FastString
consFS c fs = mkFastString (c : unpackFS fs)

461 462
uniqueOfFS :: FastString -> FastInt
uniqueOfFS (FastString u _ _ _ _) = iUnbox u
463

464
nilFS :: FastString
465
nilFS = mkFastString ""
466

467 468 469 470 471 472
-- -----------------------------------------------------------------------------
-- Stats

getFastStringTable :: IO [[FastString]]
getFastStringTable = do
  tbl <- readIORef string_table
473
  buckets <- mapM (lookupTbl tbl) [0 .. hASH_TBL_SIZE]
474 475
  return buckets

476 477 478
-- -----------------------------------------------------------------------------
-- Outputting 'FastString's

479 480
-- |Outputs a 'FastString' with /no decoding at all/, that is, you
-- get the actual bytes in the 'FastString' written to the 'Handle'.
481
hPutFS :: Handle -> FastString -> IO ()
482 483 484
hPutFS handle (FastString _ len _ fp _)
  | len == 0  = return ()
  | otherwise = do withForeignPtr fp $ \ptr -> hPutBuf handle ptr len
485

486 487
-- ToDo: we'll probably want an hPutFSLocal, or something, to output
-- in the current locale's encoding (for error messages and suchlike).
488

489 490
-- -----------------------------------------------------------------------------
-- LitStrings, here for convenience only.
491

492 493 494 495 496 497 498
-- hmm, not unboxed (or rather FastPtr), interesting
--a.k.a. Ptr CChar, Ptr Word8, Ptr (), hmph.  We don't
--really care about C types in naming, where we can help it.
type LitString = Ptr Word8
--Why do we recalculate length every time it's requested?
--If it's commonly needed, we should perhaps have
--data LitString = LitString {-#UNPACK#-}!(FastPtr Word8) {-#UNPACK#-}!FastInt
499

500
#if defined(__GLASGOW_HASKELL__)
501
mkLitString# :: Addr# -> LitString
502
mkLitString# a# = Ptr a#
503
#else
504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524
--can/should we use FastTypes here?
--Is this likely to be memory-preserving if only used on constant strings?
--should we inline it? If lucky, that would make a CAF that wouldn't
--be computationally repeated... although admittedly we're not
--really intending to use mkLitString when __GLASGOW_HASKELL__...
--(I wonder, is unicode / multi-byte characters allowed in LitStrings
-- at all?)
{-# INLINE mkLitString #-}
mkLitString :: String -> LitString
mkLitString s =
 unsafePerformIO (do
   p <- mallocBytes (length s + 1)
   let
     loop :: Int -> String -> IO ()
     loop n cs | n `seq` null cs = pokeByteOff p n (0 :: Word8)
     loop n (c:cs) = do
        pokeByteOff p n (fromIntegral (ord c) :: Word8)
        loop (1+n) cs
   loop 0 s
   return p
 )
525
#endif
526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552

unpackLitString :: LitString -> String
unpackLitString p_ = case pUnbox p_ of
 p -> unpack (_ILIT(0))
  where
    unpack n = case indexWord8OffFastPtrAsFastChar p n of
      ch -> if ch `eqFastChar` _CLIT('\0')
            then [] else cBox ch : unpack (n +# _ILIT(1))

strLength :: LitString -> Int
strLength = ptrStrLength

-- for now, use a simple String representation
--no, let's not do that right now - it's work in other places
#if 0
type LitString = String

mkLitString :: String -> LitString
mkLitString = id

unpackLitString :: LitString -> String
unpackLitString = id

strLength :: LitString -> Int
strLength = length

#endif
553 554 555 556

-- -----------------------------------------------------------------------------
-- under the carpet

Ian Lynagh's avatar
Ian Lynagh committed
557
foreign import ccall unsafe "ghc_strlen"
558
  ptrStrLength :: Ptr Word8 -> Int
559

560
-- NB. does *not* add a '\0'-terminator.
561 562
-- We only use CChar here to be parallel to the imported
-- peekC(A)StringLen.
563 564 565
pokeCAString :: Ptr CChar -> String -> IO ()
pokeCAString ptr str =
  let
566
        go [] _     = return ()
Ian Lynagh's avatar
Ian Lynagh committed
567
        go (c:cs) n = do pokeElemOff ptr n (castCharToCChar c); go cs (n+1)
568 569 570
  in
  go str 0

571
#if defined(__GLASGOW_HASKELL__) && __GLASGOW_HASKELL__ <= 602
572 573
peekCAStringLen = peekCStringLen
#endif
574
\end{code}