X-Git-Url: http://git.megacz.com/?a=blobdiff_plain;f=compiler%2Fcmm%2FCmmLex.x;h=ec9f5855c48e0719469eba6738336aff1cd08225;hb=8a2809c29de9f23eba7ca682b48390033a9d40f6;hp=c2efd17710110a278b04cfef59182ed2c28742f4;hpb=0065d5ab628975892cea1ec7303f968c3338cbe1;p=ghc-hetmet.git

diff --git a/compiler/cmm/CmmLex.x b/compiler/cmm/CmmLex.x
index c2efd17..ec9f585 100644
--- a/compiler/cmm/CmmLex.x
+++ b/compiler/cmm/CmmLex.x
@@ -1,5 +1,6 @@
 -----------------------------------------------------------------------------
--- (c) The University of Glasgow, 2004
+--
+-- (c) The University of Glasgow, 2004-2006
 --
 -- Lexer for concrete Cmm.  We try to stay close to the C-- spec, but there
 -- are a few minor differences:
@@ -24,28 +25,28 @@ import UniqFM
 import StringBuffer
 import FastString
 import Ctype
-import Util		( readRational )
+import Util
 --import TRACE
 }
 
-$whitechar   = [\ \t\n\r\f\v\xa0]
+$whitechar   = [\ \t\n\r\f\v\xa0] -- \xa0 is Unicode no-break space
 $white_no_nl = $whitechar # \n
 
 $ascdigit  = 0-9
-$unidigit  = \x01
+$unidigit  = \x01 -- Trick Alex into handling Unicode. See alexGetChar.
 $digit     = [$ascdigit $unidigit]
 $octit	   = 0-7
 $hexit     = [$digit A-F a-f]
 
-$unilarge  = \x03
+$unilarge  = \x03 -- Trick Alex into handling Unicode. See alexGetChar.
 $asclarge  = [A-Z \xc0-\xd6 \xd8-\xde]
 $large     = [$asclarge $unilarge]
 
-$unismall  = \x04
+$unismall  = \x04 -- Trick Alex into handling Unicode. See alexGetChar.
 $ascsmall  = [a-z \xdf-\xf6 \xf8-\xff]
 $small     = [$ascsmall $unismall \_]
 
-$namebegin = [$large $small \_ \. \$ \@]
+$namebegin = [$large $small \. \$ \@]
 $namechar  = [$namebegin $digit]
 
 @decimal     = $digit+
@@ -55,7 +56,7 @@ $namechar  = [$namebegin $digit]
 
 @floating_point = @decimal \. @decimal @exponent? | @decimal @exponent
 
-@escape      = \\ ([abfnrt\\\'\"\?] | x @hexadecimal | @octal)
+@escape      = \\ ([abfnrt\\\'\"\?] | x $hexit{1,2} | $octit{1,3})
 @strchar     = ($printable # [\"\\]) | @escape
 
 cmm :-
@@ -137,6 +138,8 @@ data CmmToken
   | CmmT_if
   | CmmT_jump
   | CmmT_foreign
+  | CmmT_prim
+  | CmmT_return
   | CmmT_import
   | CmmT_switch
   | CmmT_case
@@ -178,7 +181,7 @@ global_regN :: (Int -> GlobalReg) -> Action
 global_regN con span buf len 
   = return (L span (CmmT_GlobalReg (con (fromIntegral n))))
   where buf' = stepOn buf
-	n = parseInteger buf' (len-1) 10 octDecDigit
+	n = parseUnsignedInteger buf' (len-1) 10 octDecDigit
 
 global_reg :: GlobalReg -> Action
 global_reg r span buf len = return (L span (CmmT_GlobalReg r))
@@ -211,6 +214,8 @@ reservedWordsFM = listToUFM $
 	( "if",			CmmT_if ),
 	( "jump",		CmmT_jump ),
 	( "foreign",		CmmT_foreign ),
+	( "prim",		CmmT_prim ),
+	( "return",		CmmT_return ),
 	( "import",		CmmT_import ),
 	( "switch",		CmmT_switch ),
 	( "case",		CmmT_case ),
@@ -224,13 +229,13 @@ reservedWordsFM = listToUFM $
 	]
 
 tok_decimal span buf len 
-  = return (L span (CmmT_Int  $! parseInteger buf len 10 octDecDigit))
+  = return (L span (CmmT_Int  $! parseUnsignedInteger buf len 10 octDecDigit))
 
 tok_octal span buf len 
-  = return (L span (CmmT_Int  $! parseInteger (offsetBytes 1 buf) (len-1) 8 octDecDigit))
+  = return (L span (CmmT_Int  $! parseUnsignedInteger (offsetBytes 1 buf) (len-1) 8 octDecDigit))
 
 tok_hexadecimal span buf len 
-  = return (L span (CmmT_Int  $! parseInteger (offsetBytes 2 buf) (len-2) 16 hexDigit))
+  = return (L span (CmmT_Int  $! parseUnsignedInteger (offsetBytes 2 buf) (len-2) 16 hexDigit))
 
 tok_float str = CmmT_Float $! readRational str
 
@@ -242,7 +247,7 @@ tok_string str = CmmT_String (read str)
 
 setLine :: Int -> Action
 setLine code span buf len = do
-  let line = parseInteger buf len 10 octDecDigit
+  let line = parseUnsignedInteger buf len 10 octDecDigit
   setSrcLoc (mkSrcLoc (srcSpanFile span) (fromIntegral line - 1) 0)
 	-- subtract one: the line number refers to the *following* line
   -- trace ("setLine "  ++ show line) $ do
@@ -274,7 +279,7 @@ lexToken = do
   sc <- getLexState
   case alexScan inp sc of
     AlexEOF -> do let span = mkSrcSpan loc1 loc1
-		  setLastToken span 0
+		  setLastToken span 0 0
 		  return (L span CmmT_EOF)
     AlexError (loc2,_) -> do failLocMsgP loc1 loc2 "lexical error"
     AlexSkip inp2 _ -> do
@@ -283,7 +288,7 @@ lexToken = do
     AlexToken inp2@(end,buf2) len t -> do
 	setInput inp2
 	let span = mkSrcSpan loc1 end
-	span `seq` setLastToken span len
+	span `seq` setLastToken span len len
 	t span buf len
 
 -- -----------------------------------------------------------------------------