[bitc-dev] lexer bug with zero length strings

[Sam Mason]

Hi

Just found a little typo in Bitc's lexer.  It only recognises strings
with more than one character in them as valid without this patch.


  Sam
-------------- next part --------------
--- src.coyotos/src/ccs/bitcc-bootstrap/frontend/BitcLexer.l	2006-05-15 16:47:11.000000000 +0100
+++ src.coyotos/src/ccs/bitcc-bootstrap/frontend/BitcLexer.l	2006-05-15 16:47:11.000000000 +0100
@@ -329,7 +329,7 @@
 			     // But let processing continue...
 			   }
                            TOK(tk_Char);               }
-\"{SCHAR}+\"            {  unsigned badpos = validate_string(yytext);
+\"{SCHAR}*\"            {  unsigned badpos = validate_string(yytext);
                            if (badpos) {
 			     LexLoc badHere = here;
 			     badHere.offset += badpos;