path: root/ghc/compiler/utils/FastString.lhs

%
% (c) The GRASP/AQUA Project, Glasgow University, 1997-1998
%
\section{Fast strings}

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

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

Use LitString unless you want the facilities of FastString

\begin{code}
module FastString
       (
	FastString(..),     -- not abstract, for now.

        mkFastString,       -- :: String -> FastString
        mkFastStringNarrow, -- :: String -> FastString
        mkFastSubString,    -- :: Addr -> Int -> Int -> FastString

	mkFastString#,      -- :: Addr# -> FastString
        mkFastSubStringBA#, -- :: ByteArray# -> Int# -> Int# -> FastString

        mkFastStringInt,    -- :: [Int] -> FastString

        uniqueOfFS,	    -- :: FastString -> Int#
	lengthFS,	    -- :: FastString -> Int
	nullFastString,     -- :: FastString -> Bool

	unpackFS,	    -- :: FastString -> String
	unpackIntFS,	    -- :: FastString -> [Int]
	appendFS,	    -- :: FastString -> FastString -> FastString
        headFS,		    -- :: FastString -> Char
        headIntFS,	    -- :: FastString -> Int
        tailFS,		    -- :: FastString -> FastString
	concatFS,	    -- :: [FastString] -> FastString
        consFS,             -- :: Char -> FastString -> FastString
	indexFS,	    -- :: FastString -> Int -> Char
	nilFS,		    -- :: FastString

        hPutFS,		    -- :: Handle -> FastString -> IO ()

	LitString, 
	mkLitString#	    -- :: Addr# -> LitString
       ) where

-- This #define suppresses the "import FastString" that
-- HsVersions otherwise produces
#define COMPILING_FAST_STRING
#include "HsVersions.h"

#if __GLASGOW_HASKELL__ < 503
import PrelIOBase	( IO(..) )
#else
import GHC.IOBase	( IO(..) )
#endif

import PrimPacked
import GLAEXTS
import UNSAFE_IO	( unsafePerformIO )
import MONAD_ST		( stToIO )
import DATA_IOREF	( IORef, newIORef, readIORef, writeIORef )

#if __GLASGOW_HASKELL__ < 503
import PrelArr		( STArray(..), newSTArray )
#else
import GHC.Arr		( STArray(..), newSTArray )
#endif

#if __GLASGOW_HASKELL__ >= 504
import GHC.IOBase
import GHC.Handle
import Foreign.C
#else
import IOExts		( hPutBufBAFull )
#endif

import IO
import Char             ( chr, ord )

#define hASH_TBL_SIZE  4091
\end{code} 

@FastString@s are packed representations of strings
with a unique id for fast comparisons. The unique id
is assigned when creating the @FastString@, using
a hash table to map from the character string representation
to the unique ID.

\begin{code}
data FastString
  = FastString   -- packed repr. on the heap.
      Int#       -- unique id
		 --  0 => string literal, comparison
		 --  will
      Int#       -- length
      ByteArray# -- stuff

  | UnicodeStr   -- if contains characters outside '\1'..'\xFF'
      Int#       -- unique id
      [Int]      -- character numbers

instance Eq FastString where
	-- shortcut for real FastStrings
  (FastString u1 _ _) == (FastString u2 _ _) = u1 ==# u2
  a == b = case cmpFS a b of { LT -> False; EQ -> True;  GT -> False }

  (FastString u1 _ _) /= (FastString u2 _ _) = u1 /=# u2
  a /= b = case cmpFS a b of { LT -> True;  EQ -> False; GT -> True  }

instance Ord FastString where
	-- Compares lexicographically, not by unique
    a <= b = case cmpFS a b of { LT -> True;  EQ -> True;  GT -> False }
    a <	 b = case cmpFS a b of { LT -> True;  EQ -> False; GT -> False }
    a >= b = case cmpFS a b of { LT -> False; EQ -> True;  GT -> True  }
    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
    compare a b = cmpFS a b

instance Show FastString where
   show fs = show (unpackFS fs)

lengthFS :: FastString -> Int
lengthFS (FastString _ l# _) = I# l#
lengthFS (UnicodeStr _ s) = length s

nullFastString :: FastString -> Bool
nullFastString (FastString _ l# _) = l# ==# 0#
nullFastString (UnicodeStr _ []) = True
nullFastString (UnicodeStr _ (_:_)) = False

unpackFS :: FastString -> String
unpackFS (FastString _ l# ba#) = unpackNBytesBA (BA ba#) (I# l#)
unpackFS (UnicodeStr _ s) = map chr s

unpackIntFS :: FastString -> [Int]
unpackIntFS (UnicodeStr _ s) = s
unpackIntFS fs = map ord (unpackFS fs)

appendFS :: FastString -> FastString -> FastString
appendFS fs1 fs2 = mkFastStringInt (unpackIntFS fs1 ++ unpackIntFS fs2)

concatFS :: [FastString] -> FastString
concatFS ls = mkFastStringInt (concat (map unpackIntFS ls)) -- ToDo: do better

headFS :: FastString -> Char
headFS (FastString _ l# ba#) = 
 if l# ># 0# then C# (indexCharArray# ba# 0#) else error ("headFS: empty FS")
headFS (UnicodeStr _ (c:_)) = chr c
headFS (UnicodeStr _ []) = error ("headFS: empty FS")

headIntFS :: FastString -> Int
headIntFS (UnicodeStr _ (c:_)) = c
headIntFS fs = ord (headFS fs)

indexFS :: FastString -> Int -> Char
indexFS f i@(I# i#) =
 case f of
   FastString _ l# ba#
     | l# ># 0# && l# ># i#  -> C# (indexCharArray# ba# i#)
     | otherwise	     -> error (msg (I# l#))
   UnicodeStr _ s	     -> chr (s!!i)
 where
  msg l =  "indexFS: out of range: " ++ show (l,i)

tailFS :: FastString -> FastString
tailFS (FastString _ l# ba#) = mkFastSubStringBA# ba# 1# (l# -# 1#)
tailFS fs = mkFastStringInt (tail (unpackIntFS fs))

consFS :: Char -> FastString -> FastString
consFS c fs = mkFastStringInt (ord c : unpackIntFS fs)

uniqueOfFS :: FastString -> Int#
uniqueOfFS (FastString u# _ _) = u#
uniqueOfFS (UnicodeStr u# _) = u#

nilFS = mkFastString ""
\end{code}

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.

Caution: mkFastStringUnicode assumes that if the string is in the
table, it sits under the UnicodeStr constructor. Other mkFastString
variants analogously assume the FastString constructor.

\begin{code}
data FastStringTable = 
 FastStringTable
    Int#
    (MutableArray# RealWorld [FastString])

type FastStringTableVar = IORef FastStringTable

string_table :: FastStringTableVar
string_table = 
 unsafePerformIO (
   stToIO (newSTArray (0::Int,hASH_TBL_SIZE) [])
	>>= \ (STArray _ _ arr#) ->
   newIORef (FastStringTable 0# arr#))

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

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

mkFastString# :: Addr# -> FastString
mkFastString# a# =
 case strLength (Ptr a#) of { (I# len#) -> mkFastStringLen# a# len# }

mkFastStringLen# :: Addr# -> Int# -> FastString
mkFastStringLen# a# len# =
 unsafePerformIO  (
  readIORef string_table	>>= \ ft@(FastStringTable uid# tbl#) ->
  let
   h = hashStr a# len#
  in
--  _trace ("hashed: "++show (I# h)) $
  lookupTbl ft h	>>= \ lookup_result ->
  case lookup_result of
    [] -> 
       -- no match, add it to table by copying out the
       -- the string into a ByteArray
       -- _trace "empty bucket" $
       case copyPrefixStr a# (I# len#) of
	 BA barr# ->  
	   let f_str = FastString uid# len# barr# in
           updTbl string_table ft h [f_str] >>
           ({- _trace ("new: " ++ show f_str)   $ -} return f_str)
    ls -> 
       -- non-empty `bucket', scan the list looking
       -- entry with same length and compare byte by byte.
       -- _trace ("non-empty bucket"++show ls) $
       case bucket_match ls len# a# of
	 Nothing -> 
           case copyPrefixStr a# (I# len#) of
	     BA barr# ->  
              let f_str = FastString uid# len# barr# in
              updTbl string_table ft h (f_str:ls) >>
	      ( {- _trace ("new: " ++ show f_str)  $ -} return f_str)
	 Just v  -> {- _trace ("re-use: "++show v) $ -} return v)
  where
   bucket_match [] _ _ = Nothing
   bucket_match (v@(FastString _ l# ba#):ls) len# a# =
      if len# ==# l# && eqStrPrefix a# ba# l# then
	 Just v
      else
	 bucket_match ls len# a#
   bucket_match (UnicodeStr _ _ : ls) len# a# =
      bucket_match ls len# a#

mkFastSubStringBA# :: ByteArray# -> Int# -> Int# -> FastString
mkFastSubStringBA# barr# start# len# =
 unsafePerformIO  (
  readIORef string_table	>>= \ ft@(FastStringTable uid# tbl#) ->
  let
   h = hashSubStrBA barr# start# len#
  in
--  _trace ("hashed(b): "++show (I# h)) $
  lookupTbl ft h		>>= \ lookup_result ->
  case lookup_result of
    [] -> 
       -- no match, add it to table by copying out the
       -- the string into a ByteArray
       -- _trace "empty bucket(b)" $
       case copySubStrBA (BA barr#) (I# start#) (I# len#) of
         BA ba# ->  
          let f_str = FastString uid# len# ba# in
          updTbl string_table ft h [f_str]     >>
          -- _trace ("new(b): " ++ show f_str)   $
	  return f_str
    ls ->
       -- non-empty `bucket', scan the list looking
       -- entry with same length and compare byte by byte. 
       -- _trace ("non-empty bucket(b)"++show ls) $
       case bucket_match ls start# len# barr# of
	 Nothing ->
          case copySubStrBA (BA barr#) (I# start#) (I# len#) of
            BA ba# ->  
              let f_str = FastString uid# len# ba# in
              updTbl string_table ft h (f_str:ls) >>
	      -- _trace ("new(b): " ++ show f_str)   $
	      return f_str
	 Just v  -> 
              -- _trace ("re-use(b): "++show v) $
	      return v
  )
 where
   bucket_match [] _ _ _ = Nothing
   bucket_match (v:ls) start# len# ba# =
    case v of
     FastString _ l# barr# ->
      if len# ==# l# && eqStrPrefixBA barr# ba# start# len# then
	 Just v
      else
	 bucket_match ls start# len# ba#
     UnicodeStr _ _ -> bucket_match ls start# len# ba#

mkFastStringUnicode :: [Int] -> FastString
mkFastStringUnicode s =
 unsafePerformIO  (
  readIORef string_table	>>= \ ft@(FastStringTable uid# tbl#) ->
  let
   h = hashUnicode s 0#
  in
--  _trace ("hashed(b): "++show (I# h)) $
  lookupTbl ft h		>>= \ lookup_result ->
  case lookup_result of
    [] -> 
       -- no match, add it to table by copying out the
       -- the string into a [Int]
          let f_str = UnicodeStr uid# s in
          updTbl string_table ft h [f_str]     >>
          -- _trace ("new(b): " ++ show f_str)   $
	  return f_str
    ls -> 
       -- non-empty `bucket', scan the list looking
       -- entry with same length and compare byte by byte. 
       -- _trace ("non-empty bucket(b)"++show ls) $
       case bucket_match ls of
	 Nothing -> 
              let f_str = UnicodeStr uid# s in
              updTbl string_table ft h (f_str:ls) >>
	      -- _trace ("new(b): " ++ show f_str)   $
	      return f_str
	 Just v  -> 
              -- _trace ("re-use(b): "++show v) $
	      return v
  )
 where
   bucket_match [] = Nothing
   bucket_match (v@(UnicodeStr _ s'):ls) =
       if s' == s then Just v else bucket_match ls
   bucket_match (FastString _ _ _ : ls) = bucket_match ls

mkFastStringNarrow :: String -> FastString
mkFastStringNarrow str =
 case packString str of { (I# len#, BA frozen#) -> 
    mkFastSubStringBA# frozen# 0# len#
 }
 {- 0-indexed array, len# == index to one beyond end of string,
    i.e., (0,1) => empty string.    -}

mkFastString :: String -> FastString
mkFastString str = if all good str
    then mkFastStringNarrow str
    else mkFastStringUnicode (map ord str)
    where
    good c = c >= '\1' && c <= '\xFF'

mkFastStringInt :: [Int] -> FastString
mkFastStringInt str = if all good str
    then mkFastStringNarrow (map chr str)
    else mkFastStringUnicode str
    where
    good c = c >= 1 && c <= 0xFF

mkFastSubString :: Addr# -> Int -> Int -> FastString
mkFastSubString a# (I# start#) (I# len#) =
 mkFastStringLen# (a# `plusAddr#` start#) len#
\end{code}

\begin{code}
hashStr  :: Addr# -> Int# -> Int#
 -- use the Addr to produce a hash value between 0 & m (inclusive)
hashStr a# len# = loop 0# 0#
   where 
    loop h n | n ==# len# = h
	     | otherwise  = loop h2 (n +# 1#)
	  where c = ord# (indexCharOffAddr# a# n)
		h2 = (c +# (h *# 128#)) `remInt#` hASH_TBL_SIZE#

hashSubStrBA  :: ByteArray# -> Int# -> Int# -> Int#
 -- use the byte array to produce a hash value between 0 & m (inclusive)
hashSubStrBA ba# start# len# = loop 0# 0#
   where 
    loop h n | n ==# len# = h
	     | otherwise  = loop h2 (n +# 1#)
	  where c = ord# (indexCharArray# ba# (start# +# n))
		h2 = (c +# (h *# 128#)) `remInt#` hASH_TBL_SIZE#

hashUnicode :: [Int] -> Int# -> Int#
hashUnicode [] h = h
hashUnicode (I# c : cs) h = hashUnicode cs ((c +# (h *# 128#)) `remInt#` hASH_TBL_SIZE#)
\end{code}

\begin{code}
cmpFS :: FastString -> FastString -> Ordering
cmpFS (UnicodeStr u1# s1) (UnicodeStr u2# s2) = if u1# ==# u2# then EQ
    else compare s1 s2
cmpFS (UnicodeStr _ s1) s2 = compare s1 (unpackIntFS s2)
cmpFS s1 (UnicodeStr _ s2) = compare (unpackIntFS s1) s2
cmpFS (FastString u1# l1# b1#) (FastString u2# l2# b2#) =
  if u1# ==# u2# then EQ else
  let l# = if l1# <=# l2# then l1# else l2# in
  unsafePerformIO (
    memcmp b1# b2# l# >>= \ (I# res) ->
    return (
    if      res <#  0# then LT
    else if res ==# 0# then 
	if l1# ==# l2# then EQ
	else if l1# <# l2# then LT else GT
    else		    GT
    ))

#ifndef __HADDOCK__
foreign import ccall unsafe "ghc_memcmp" 
  memcmp :: ByteArray# -> ByteArray# -> Int# -> IO Int
#endif

-- -----------------------------------------------------------------------------
-- Outputting 'FastString's

#if __GLASGOW_HASKELL__ >= 504

-- this is our own version of hPutBuf for FastStrings, because in
-- 5.04+ we don't have mutable byte arrays and therefore hPutBufBA.
-- The closest is hPutArray in Data.Array.IO, but that does some extra
-- range checks that we want to avoid here.

foreign import ccall unsafe "__hscore_memcpy_dst_off"
   memcpy_baoff_ba :: RawBuffer -> Int -> RawBuffer -> CSize -> IO (Ptr ())

hPutFS handle (FastString _ l# ba#)
  | l# ==# 0#  = return ()
  | otherwise
   = do wantWritableHandle "hPutFS" handle $ 
          \ handle_@Handle__{ haFD=fd, haBuffer=ref, haIsStream=stream } -> do

          old_buf@Buffer{ bufBuf=old_raw, bufRPtr=r, bufWPtr=w, bufSize=size }
	    <- readIORef ref

	  let count = I# l#
	      raw = unsafeCoerce# ba# :: MutableByteArray# RealWorld

          -- enough room in handle buffer?
          if (size - w > count)
		-- There's enough room in the buffer:
		-- just copy the data in and update bufWPtr.
	    then do memcpy_baoff_ba old_raw w raw (fromIntegral count)
		    writeIORef ref old_buf{ bufWPtr = w + count }
		    return ()

		-- else, we have to flush
	    else do flushed_buf <- flushWriteBuffer fd stream old_buf
		    writeIORef ref flushed_buf
		    let this_buf = 
			    Buffer{ bufBuf=raw, bufState=WriteBuffer, 
				    bufRPtr=0, bufWPtr=count, bufSize=count }
		    flushWriteBuffer fd stream this_buf
		    return ()

#else

hPutFS :: Handle -> FastString -> IO ()
hPutFS handle (FastString _ l# ba#)
  | l# ==# 0#  = return ()
  | otherwise  = do mba <- stToIO $ unsafeThawByteArray (ByteArray (bot::Int) bot ba#)
                    hPutBufBAFull  handle mba (I# l#)
 where
  bot = error "hPutFS.ba"

#endif

-- ONLY here for debugging the NCG (so -ddump-stix works for string
-- literals); no idea if this is really necessary.  JRS, 010131
hPutFS handle (UnicodeStr _ is) 
  = hPutStr handle ("(UnicodeStr " ++ show is ++ ")")

-- -----------------------------------------------------------------------------
-- LitStrings, here for convenience only.

type LitString = Ptr ()

mkLitString# :: Addr# -> LitString
mkLitString# a# = Ptr a#
\end{code}