support 32 bit runes

Reference: /n/sources/patch/saved/bloated-rune-cc
Date: Fri Jan 29 07:00:26 CET 2010
Signed-off-by: quanstro@quanstro.net

--- /sys/src/cmd/cc/com.c	Fri Jan 29 06:59:20 2010
+++ /sys/src/cmd/cc/com.c	Fri Jan 29 06:59:18 2010
@@ -617,10 +617,10 @@
 		break;
 
 	case OLSTRING:
-		if(n->type->link != types[TUSHORT]) {
+		if(n->type->link != types[TRUNE]) {
 			o = outstring(0, 0);
 			while(o & 3) {
-				outlstring(L"", sizeof(ushort));
+				outlstring(L"", sizeof(Rune));
 				o = outlstring(0, 0);
 			}
 		}
--- /sys/src/cmd/cc/cc.y	Fri Jan 29 06:59:28 2010
+++ /sys/src/cmd/cc/cc.y	Fri Jan 29 06:59:26 2010
@@ -853,9 +853,9 @@
 	LLSTRING
 	{
 		$$ = new(OLSTRING, Z, Z);
-		$$->type = typ(TARRAY, types[TUSHORT]);
-		$$->type->width = $1.l + sizeof(ushort);
-		$$->rstring = (ushort*)$1.s;
+		$$->type = typ(TARRAY, types[TRUNE]);
+		$$->type->width = $1.l + sizeof(Rune);
+		$$->rstring = (Rune*)$1.s;
 		$$->sym = symstring;
 		$$->etype = TARRAY;
 		$$->class = CSTATIC;
@@ -865,16 +865,16 @@
 		char *s;
 		int n;
 
-		n = $1->type->width - sizeof(ushort);
+		n = $1->type->width - sizeof(Rune);
 		s = alloc(n+$2.l+MAXALIGN);
 
 		memcpy(s, $1->rstring, n);
 		memcpy(s+n, $2.s, $2.l);
-		*(ushort*)(s+n+$2.l) = 0;
+		*(Rune*)(s+n+$2.l) = 0;
 
 		$$ = $1;
 		$$->type->width += $2.l;
-		$$->rstring = (ushort*)s;
+		$$->rstring = (Rune*)s;
 	}
 
 zelist:
--- /sys/src/cmd/cc/pswt.c	Fri Jan 29 06:59:35 2010
+++ /sys/src/cmd/cc/pswt.c	Fri Jan 29 06:59:34 2010
@@ -78,10 +78,10 @@
 }
 
 long
-outlstring(ushort *s, long n)
+outlstring(Rune *s, long n)
 {
-	char buf[2];
-	int c;
+	char buf[UTFmax];
+	int c, i;
 	long r;
 
 	if(suppress)
@@ -92,14 +92,14 @@
 	while(n > 0) {
 		c = *s++;
 		if(align(0, types[TCHAR], Aarg1)) {
-			buf[0] = c>>8;
-			buf[1] = c;
+			for(i = 0; i < sizeof buf; i++)
+				buf[i] = c>>8*(sizeof buf - i - 1);
 		} else {
-			buf[0] = c;
-			buf[1] = c>>8;
+			for(i = 0; i < sizeof buf; i++)
+				buf[i] = c>>8*i;
 		}
-		outstring(buf, 2);
-		n -= sizeof(ushort);
+		outstring(buf, sizeof(Rune));
+		n -= sizeof(Rune);
 	}
 	return r;
 }
--- /sys/src/cmd/cc/cc.h	Fri Jan 29 06:59:44 2010
+++ /sys/src/cmd/cc/cc.h	Fri Jan 29 06:59:43 2010
@@ -51,7 +51,7 @@
 	double	fconst;		/* fp constant */
 	vlong	vconst;		/* non fp const */
 	char*	cstring;	/* character string */
-	ushort*	rstring;	/* rune string */
+	Rune*	rstring;	/* rune string */
 
 	Sym*	sym;
 	Type*	type;
@@ -335,6 +335,12 @@
 	TFILE,
 	TOLD,
 	NALLTYPES,
+
+	/*
+	 * bootstrapping
+	 */
+//	TRUNE = TUINT,
+	TRUNE = TUSHORT + sizeof(Rune)==4,
 };
 enum
 {
@@ -735,7 +741,7 @@
 void	gextern(Sym*, Node*, long, long);
 void	ginit(void);
 long	outstring(char*, long);
-long	outlstring(ushort*, long);
+long	outlstring(Rune*, long);
 void	sextern(Sym*, Node*, long, long);
 void	xcom(Node*);
 long	exreg(Type*);
--- /sys/src/cmd/cc/lex.c	Fri Jan 29 06:59:56 2010
+++ /sys/src/cmd/cc/lex.c	Fri Jan 29 06:59:54 2010
@@ -469,7 +469,7 @@
 				yyerror("missing '");
 				peekc = c1;
 			}
-			yylval.vval = convvtox(c, TUSHORT);
+			yylval.vval = convvtox(c, TRUNE);
 			return LUCONST;
 		}
 		if(c == '"') {
@@ -543,15 +543,15 @@
 			c = escchar('"', 1, 0);
 			if(c == EOF)
 				break;
-			cp = allocn(cp, c1, sizeof(ushort));
-			*(ushort*)(cp + c1) = c;
-			c1 += sizeof(ushort);
+			cp = allocn(cp, c1, sizeof(Rune));
+			*(Rune*)(cp + c1) = c;
+			c1 += sizeof(Rune);
 		}
 		yylval.sval.l = c1;
 		do {
-			cp = allocn(cp, c1, sizeof(ushort));
-			*(ushort*)(cp + c1) = 0;
-			c1 += sizeof(ushort);
+			cp = allocn(cp, c1, sizeof(Rune));
+			*(Rune*)(cp + c1) = 0;
+			c1 += sizeof(Rune);
 		} while(c1 & MAXALIGN);
 		yylval.sval.s = cp;
 		return LLSTRING;
@@ -1029,7 +1029,7 @@
 	} else
 		c = GETC();
 	for(;;) {
-		if(!isspace(c))
+		if(c >= Runeself || !isspace(c))
 			return c;
 		if(c == '\n') {
 			lineno++;