FastString.lhs 18.1 KB
Newer Older
1
%
2
% (c) The University of Glasgow, 1997-2006
3
%
4
\begin{code}
5 6 7 8
{-# OPTIONS -fno-warn-unused-imports #-}
-- XXX GHC 6.9 seems to be confused by unpackCString# being used only in
--     a RULE

9 10 11 12
{-# OPTIONS_GHC -O -funbox-strict-fields #-}
-- We always optimise this, otherwise performance of a non-optimised
-- compiler is severely affected

13
{-
Ian Lynagh's avatar
Ian Lynagh committed
14 15
FastString:     A compact, hash-consed, representation of character strings.
                Comparison is O(1), and you can get a Unique from them.
16
                Generated by fsLit
Ian Lynagh's avatar
Ian Lynagh committed
17
                Turn into SDoc with Outputable.ftext
18

Ian Lynagh's avatar
Ian Lynagh committed
19 20 21
LitString:      Just a wrapper for the Addr# of a C string (Ptr CChar).
                Practically no operations
                Outputing them is fast
22
                Generated by sLit
Ian Lynagh's avatar
Ian Lynagh committed
23
                Turn into SDoc with Outputable.ptext
24 25

Use LitString unless you want the facilities of FastString
26
-}
27 28
module FastString
       (
Ian Lynagh's avatar
Ian Lynagh committed
29 30
        -- * FastStrings
        FastString(..),     -- not abstract, for now.
31

Ian Lynagh's avatar
Ian Lynagh committed
32
        -- ** Construction
33
        mkFastString,
Ian Lynagh's avatar
Ian Lynagh committed
34
        mkFastStringBytes,
35
        mkFastStringByteList,
Ian Lynagh's avatar
Ian Lynagh committed
36
        mkFastStringForeignPtr,
37
#if defined(__GLASGOW_HASKELL__)
Ian Lynagh's avatar
Ian Lynagh committed
38
        mkFastString#,
39
#endif
Ian Lynagh's avatar
Ian Lynagh committed
40 41
        mkZFastString,
        mkZFastStringBytes,
42

Ian Lynagh's avatar
Ian Lynagh committed
43 44 45
        -- ** Deconstruction
        unpackFS,           -- :: FastString -> String
        bytesFS,            -- :: FastString -> [Word8]
46

Ian Lynagh's avatar
Ian Lynagh committed
47 48 49
        -- ** Encoding
        isZEncoded,
        zEncodeFS,
50

Ian Lynagh's avatar
Ian Lynagh committed
51
        -- ** Operations
52
        uniqueOfFS,
Ian Lynagh's avatar
Ian Lynagh committed
53 54 55
        lengthFS,
        nullFS,
        appendFS,
56 57
        headFS,
        tailFS,
Ian Lynagh's avatar
Ian Lynagh committed
58
        concatFS,
59
        consFS,
Ian Lynagh's avatar
Ian Lynagh committed
60
        nilFS,
61

Ian Lynagh's avatar
Ian Lynagh committed
62
        -- ** Outputing
63 64
        hPutFS,

Ian Lynagh's avatar
Ian Lynagh committed
65 66 67
        -- ** Internal
        getFastStringTable,
        hasZEncoding,
68

Ian Lynagh's avatar
Ian Lynagh committed
69 70
        -- * LitStrings
        LitString,
71
#if defined(__GLASGOW_HASKELL__)
Ian Lynagh's avatar
Ian Lynagh committed
72
        mkLitString#,
73
#endif
Ian Lynagh's avatar
Ian Lynagh committed
74
        mkLitString,
Ian Lynagh's avatar
Ian Lynagh committed
75 76
        unpackLitString,
        strLength,
77

Ian Lynagh's avatar
Ian Lynagh committed
78 79 80 81
        ptrStrLength,

        sLit,
        fsLit,
82 83
       ) where

84
#include "HsVersions.h"
85

86
import Encoding
87 88
import FastTypes
import FastFunctions
89
import Panic
90

91 92
import Foreign
import Foreign.C
93
import GHC.Exts
94
import System.IO
95
import System.IO.Unsafe ( unsafePerformIO )
Ian Lynagh's avatar
Ian Lynagh committed
96 97 98
import Data.IORef       ( IORef, newIORef, readIORef, writeIORef )
import Data.Maybe       ( isJust )
import Data.Char        ( ord )
99

Ian Lynagh's avatar
Ian Lynagh committed
100 101
import GHC.IOBase       ( IO(..) )
import GHC.Ptr          ( Ptr(..) )
Ian Lynagh's avatar
Ian Lynagh committed
102 103 104
#if defined(__GLASGOW_HASKELL__)
import GHC.Base         ( unpackCString# )
#endif
105

106 107
#define hASH_TBL_SIZE          4091
#define hASH_TBL_SIZE_UNBOXED  4091#
108 109


110 111 112 113 114
{-|
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.
115

116 117 118 119
'FastString's support a memoized conversion to the Z-encoding via zEncodeFS.
-}

data FastString = FastString {
Ian Lynagh's avatar
Ian Lynagh committed
120 121 122
      uniq    :: {-# UNPACK #-} !Int, -- unique id
      n_bytes :: {-# UNPACK #-} !Int, -- number of bytes
      n_chars :: {-# UNPACK #-} !Int, -- number of chars
123 124 125 126 127
      buf     :: {-# UNPACK #-} !(ForeignPtr Word8),
      enc     :: FSEncoding
  }

data FSEncoding
Ian Lynagh's avatar
Ian Lynagh committed
128
    -- including strings that don't need any encoding
129
  = ZEncoded
Ian Lynagh's avatar
Ian Lynagh committed
130
    -- A UTF-8 string with a memoized Z-encoding
131
  | UTF8Encoded {-# UNPACK #-} !(IORef (Maybe FastString))
132

133 134
instance Eq FastString where
  f1 == f2  =  uniq f1 == uniq f2
135

136
instance Ord FastString where
Ian Lynagh's avatar
Ian Lynagh committed
137
    -- Compares lexicographically, not by unique
138
    a <= b = case cmpFS a b of { LT -> True;  EQ -> True;  GT -> False }
Ian Lynagh's avatar
Ian Lynagh committed
139
    a <  b = case cmpFS a b of { LT -> True;  EQ -> False; GT -> False }
140
    a >= b = case cmpFS a b of { LT -> False; EQ -> True;  GT -> True  }
Ian Lynagh's avatar
Ian Lynagh committed
141 142 143 144 145
    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
146
    compare a b = cmpFS a b
147

148 149 150
instance Show FastString where
   show fs = show (unpackFS fs)

151 152 153
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
154 155 156 157 158 159 160 161 162 163 164
  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
165

166
#ifndef __HADDOCK__
Ian Lynagh's avatar
Ian Lynagh committed
167
foreign import ccall unsafe "ghc_memcmp"
168 169
  memcmp :: Ptr a -> Ptr b -> Int -> IO Int
#endif
170

171 172
-- -----------------------------------------------------------------------------
-- Construction
173

174
{-
175 176 177 178
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.
179
-}
180

Ian Lynagh's avatar
Ian Lynagh committed
181
data FastStringTable =
182
 FastStringTable
183
    {-# UNPACK #-} !Int
184
    (MutableArray# RealWorld [FastString])
185

186
{-# NOINLINE string_table #-}
187
string_table :: IORef FastStringTable
Ian Lynagh's avatar
Ian Lynagh committed
188
string_table =
189
 unsafePerformIO $ do
190 191 192 193
   tab <- IO $ \s1# -> case newArray# hASH_TBL_SIZE_UNBOXED [] s1# of
                           (# s2#, arr# #) ->
                               (# s2#, FastStringTable 0 arr# #)
   newIORef tab
194 195 196 197 198 199 200 201 202

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#)
203

204
mkFastString# :: Addr# -> FastString
205
mkFastString# a# = mkFastStringBytes ptr (ptrStrLength ptr)
206
  where ptr = Ptr a#
207

208 209
mkFastStringBytes :: Ptr Word8 -> Int -> FastString
mkFastStringBytes ptr len = unsafePerformIO $ do
210
  ft@(FastStringTable uid _) <- readIORef string_table
211
  let
212 213
   h = hashStr ptr len
   add_it ls = do
Ian Lynagh's avatar
Ian Lynagh committed
214 215 216 217
        fs <- copyNewFastString uid ptr len
        updTbl string_table ft h (fs:ls)
        {- _trace ("new: " ++ show f_str)   $ -}
        return fs
218 219
  --
  lookup_result <- lookupTbl ft h
sof's avatar
sof committed
220
  case lookup_result of
221 222 223 224
    [] -> add_it []
    ls -> do
       b <- bucket_match ls len ptr
       case b of
Ian Lynagh's avatar
Ian Lynagh committed
225 226
         Nothing -> add_it ls
         Just v  -> {- _trace ("re-use: "++show v) $ -} return v
227 228 229

mkZFastStringBytes :: Ptr Word8 -> Int -> FastString
mkZFastStringBytes ptr len = unsafePerformIO $ do
230
  ft@(FastStringTable uid _) <- readIORef string_table
sof's avatar
sof committed
231
  let
232 233
   h = hashStr ptr len
   add_it ls = do
Ian Lynagh's avatar
Ian Lynagh committed
234 235 236 237
        fs <- copyNewZFastString uid ptr len
        updTbl string_table ft h (fs:ls)
        {- _trace ("new: " ++ show f_str)   $ -}
        return fs
238 239
  --
  lookup_result <- lookupTbl ft h
sof's avatar
sof committed
240
  case lookup_result of
241 242 243 244
    [] -> add_it []
    ls -> do
       b <- bucket_match ls len ptr
       case b of
Ian Lynagh's avatar
Ian Lynagh committed
245 246
         Nothing -> add_it ls
         Just v  -> {- _trace ("re-use: "++show v) $ -} return v
247 248 249 250 251 252

-- | 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
253
  ft@(FastStringTable uid _) <- readIORef string_table
254
--  _trace ("hashed: "++show (I# h)) $
255
  let
256 257
    h = hashStr ptr len
    add_it ls = do
Ian Lynagh's avatar
Ian Lynagh committed
258 259 260 261
        fs <- mkNewFastString uid ptr fp len
        updTbl string_table ft h (fs:ls)
        {- _trace ("new: " ++ show f_str)   $ -}
        return fs
262 263 264 265 266 267 268
  --
  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
269 270
         Nothing -> add_it ls
         Just v  -> {- _trace ("re-use: "++show v) $ -} return v
271 272 273

mkZFastStringForeignPtr :: Ptr Word8 -> ForeignPtr Word8 -> Int -> IO FastString
mkZFastStringForeignPtr ptr fp len = do
274
  ft@(FastStringTable uid _) <- readIORef string_table
275 276 277 278
--  _trace ("hashed: "++show (I# h)) $
  let
    h = hashStr ptr len
    add_it ls = do
Ian Lynagh's avatar
Ian Lynagh committed
279 280 281 282
        fs <- mkNewZFastString uid ptr fp len
        updTbl string_table ft h (fs:ls)
        {- _trace ("new: " ++ show f_str)   $ -}
        return fs
283 284
  --
  lookup_result <- lookupTbl ft h
285
  case lookup_result of
286 287 288 289
    [] -> add_it []
    ls -> do
       b <- bucket_match ls len ptr
       case b of
Ian Lynagh's avatar
Ian Lynagh committed
290 291
         Nothing -> add_it ls
         Just v  -> {- _trace ("re-use: "++show v) $ -} return v
292 293


294 295
-- | Creates a UTF-8 encoded 'FastString' from a 'String'
mkFastString :: String -> FastString
Ian Lynagh's avatar
Ian Lynagh committed
296
mkFastString str =
297 298 299 300 301
  inlinePerformIO $ do
    let l = utf8EncodedLength str
    buf <- mallocForeignPtrBytes l
    withForeignPtr buf $ \ptr -> do
      utf8EncodeString ptr str
Ian Lynagh's avatar
Ian Lynagh committed
302
      mkFastStringForeignPtr ptr buf l
303

304 305
-- | Creates a 'FastString' from a UTF-8 encoded @[Word8]@
mkFastStringByteList :: [Word8] -> FastString
Ian Lynagh's avatar
Ian Lynagh committed
306
mkFastStringByteList str =
307 308 309 310 311
  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
312
      mkFastStringForeignPtr ptr buf l
313 314 315

-- | Creates a Z-encoded 'FastString' from a 'String'
mkZFastString :: String -> FastString
Ian Lynagh's avatar
Ian Lynagh committed
316
mkZFastString str =
317 318 319 320 321
  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
322
      mkZFastStringForeignPtr ptr buf l
323

324
bucket_match :: [FastString] -> Int -> Ptr Word8 -> IO (Maybe FastString)
325 326 327
bucket_match [] _ _ = return Nothing
bucket_match (v@(FastString _ l _ buf _):ls) len ptr
      | len == l  =  do
Ian Lynagh's avatar
Ian Lynagh committed
328 329 330 331 332
         b <- cmpStringPrefix ptr buf len
         if b then return (Just v)
              else bucket_match ls len ptr
      | otherwise =
         bucket_match ls len ptr
333

334 335
mkNewFastString :: Int -> Ptr Word8 -> ForeignPtr Word8 -> Int
                -> IO FastString
336 337 338 339 340
mkNewFastString uid ptr fp len = do
  ref <- newIORef Nothing
  n_chars <- countUTF8Chars ptr len
  return (FastString uid len n_chars fp (UTF8Encoded ref))

341 342 343
mkNewZFastString :: Int -> Ptr Word8 -> ForeignPtr Word8 -> Int
                 -> IO FastString
mkNewZFastString uid _ fp len = do
344 345
  return (FastString uid len len fp ZEncoded)

346
copyNewFastString :: Int -> Ptr Word8 -> Int -> IO FastString
347 348 349 350 351 352
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))

353
copyNewZFastString :: Int -> Ptr Word8 -> Int -> IO FastString
354 355 356 357
copyNewZFastString uid ptr len = do
  fp <- copyBytesToForeignPtr ptr len
  return (FastString uid len len fp ZEncoded)

358
copyBytesToForeignPtr :: Ptr Word8 -> Int -> IO (ForeignPtr Word8)
359 360 361 362 363 364 365 366 367 368 369 370 371
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
372
 -- use the Addr to produce a hash value between 0 & m (inclusive)
373
hashStr (Ptr a#) (I# len#) = loop 0# 0#
Ian Lynagh's avatar
Ian Lynagh committed
374
   where
375
    loop h n | n GHC.Exts.==# len# = I# h
Ian Lynagh's avatar
Ian Lynagh committed
376 377 378 379
             | 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#
380

381 382
-- -----------------------------------------------------------------------------
-- Operations
383

384 385 386
-- | Returns the length of the 'FastString' in characters
lengthFS :: FastString -> Int
lengthFS f = n_chars f
387

388 389 390
-- | Returns 'True' if the 'FastString' is Z-encoded
isZEncoded :: FastString -> Bool
isZEncoded fs | ZEncoded <- enc fs = True
Ian Lynagh's avatar
Ian Lynagh committed
391
              | otherwise          = False
392

393 394 395
-- | Returns 'True' if this 'FastString' is not Z-encoded but already has
-- a Z-encoding cached (used in producing stats).
hasZEncoding :: FastString -> Bool
396
hasZEncoding (FastString _ _ _ _ enc) =
397 398 399 400 401
  case enc of
    ZEncoded -> False
    UTF8Encoded ref ->
      inlinePerformIO $ do
        m <- readIORef ref
Ian Lynagh's avatar
Ian Lynagh committed
402
        return (isJust m)
403

404 405 406 407 408 409
-- | 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
410
unpackFS (FastString _ n_bytes _ buf enc) =
411 412
  inlinePerformIO $ withForeignPtr buf $ \ptr ->
    case enc of
Ian Lynagh's avatar
Ian Lynagh committed
413 414
        ZEncoded      -> peekCAStringLen (castPtr ptr,n_bytes)
        UTF8Encoded _ -> utf8DecodeString ptr n_bytes
415 416

bytesFS :: FastString -> [Word8]
417
bytesFS (FastString _ n_bytes _ buf _) =
418 419 420 421 422 423 424 425 426
  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
427
zEncodeFS fs@(FastString _ _ _ _ enc) =
428 429 430 431 432 433
  case enc of
    ZEncoded -> fs
    UTF8Encoded ref ->
      inlinePerformIO $ do
        m <- readIORef ref
        case m of
Ian Lynagh's avatar
Ian Lynagh committed
434 435
          Just fs -> return fs
          Nothing -> do
436
            let efs = mkZFastString (zEncodeString (unpackFS fs))
Ian Lynagh's avatar
Ian Lynagh committed
437 438
            writeIORef ref (Just efs)
            return efs
439 440 441 442 443 444 445 446

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
447 448
headFS (FastString _ 0 _ _ _) = panic "headFS: Empty FastString"
headFS (FastString _ _ _ buf enc) =
449 450
  inlinePerformIO $ withForeignPtr buf $ \ptr -> do
    case enc of
Ian Lynagh's avatar
Ian Lynagh committed
451 452 453 454 455
      ZEncoded -> do
         w <- peek (castPtr ptr)
         return (castCCharToChar w)
      UTF8Encoded _ ->
         return (fst (utf8DecodeChar ptr))
456 457

tailFS :: FastString -> FastString
458
tailFS (FastString _ 0 _ _ _) = panic "tailFS: Empty FastString"
Ian Lynagh's avatar
Ian Lynagh committed
459
tailFS (FastString _ n_bytes _ buf enc) =
460 461 462
  inlinePerformIO $ withForeignPtr buf $ \ptr -> do
    case enc of
      ZEncoded -> do
Ian Lynagh's avatar
Ian Lynagh committed
463
        return $! mkZFastStringBytes (ptr `plusPtr` 1) (n_bytes - 1)
464
      UTF8Encoded _ -> do
Ian Lynagh's avatar
Ian Lynagh committed
465 466 467
         let (_,ptr') = utf8DecodeChar ptr
         let off = ptr' `minusPtr` ptr
         return $! mkFastStringBytes (ptr `plusPtr` off) (n_bytes - off)
468 469 470 471

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

472 473
uniqueOfFS :: FastString -> FastInt
uniqueOfFS (FastString u _ _ _ _) = iUnbox u
474

475
nilFS :: FastString
476
nilFS = mkFastString ""
477

478 479 480 481 482 483
-- -----------------------------------------------------------------------------
-- Stats

getFastStringTable :: IO [[FastString]]
getFastStringTable = do
  tbl <- readIORef string_table
484
  buckets <- mapM (lookupTbl tbl) [0 .. hASH_TBL_SIZE]
485 486
  return buckets

487 488 489
-- -----------------------------------------------------------------------------
-- Outputting 'FastString's

490 491
-- |Outputs a 'FastString' with /no decoding at all/, that is, you
-- get the actual bytes in the 'FastString' written to the 'Handle'.
492
hPutFS :: Handle -> FastString -> IO ()
493 494 495
hPutFS handle (FastString _ len _ fp _)
  | len == 0  = return ()
  | otherwise = do withForeignPtr fp $ \ptr -> hPutBuf handle ptr len
496

497 498
-- ToDo: we'll probably want an hPutFSLocal, or something, to output
-- in the current locale's encoding (for error messages and suchlike).
499

500 501
-- -----------------------------------------------------------------------------
-- LitStrings, here for convenience only.
502

503 504 505 506 507 508 509
-- 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
510

511
#if defined(__GLASGOW_HASKELL__)
512
mkLitString# :: Addr# -> LitString
513
mkLitString# a# = Ptr a#
Ian Lynagh's avatar
Ian Lynagh committed
514
#endif
515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
--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
Ian Lynagh's avatar
Ian Lynagh committed
533 534 535
     -- XXX GHC isn't smart enough to know that we have already covered
     -- this case.
     loop _ [] = panic "mkLitString"
536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565
   loop 0 s
   return p
 )

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
566 567 568 569

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

Ian Lynagh's avatar
Ian Lynagh committed
570
foreign import ccall unsafe "ghc_strlen"
571
  ptrStrLength :: Ptr Word8 -> Int
572

573
-- NB. does *not* add a '\0'-terminator.
574 575
-- We only use CChar here to be parallel to the imported
-- peekC(A)StringLen.
576 577 578
pokeCAString :: Ptr CChar -> String -> IO ()
pokeCAString ptr str =
  let
579
        go [] _     = return ()
Ian Lynagh's avatar
Ian Lynagh committed
580
        go (c:cs) n = do pokeElemOff ptr n (castCharToCChar c); go cs (n+1)
581 582 583
  in
  go str 0

Ian Lynagh's avatar
Ian Lynagh committed
584 585 586 587 588 589 590 591 592 593 594 595
{-# NOINLINE sLit #-}
sLit :: String -> LitString
sLit x  = mkLitString x

{-# NOINLINE fsLit #-}
fsLit :: String -> FastString
fsLit x = mkFastString x

{-# RULES "slit"
    forall x . sLit  (unpackCString# x) = mkLitString#  x #-}
{-# RULES "fslit"
    forall x . fsLit (unpackCString# x) = mkFastString# x #-}
596
\end{code}