fix utf and rune handling in preparation for 32bit runes

6 files changed, 38 insertions, 31 deletions

diff --git a/sys/src/cmd/cc/cc.h b/sys/src/cmd/cc/cc.h
index d66faaa11..01eb04562 100644
--- a/sys/src/cmd/cc/cc.h
+++ b/sys/src/cmd/cc/cc.h
@@ -51,7 +51,7 @@ struct	Node
 	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;
@@ -336,6 +336,8 @@ enum
 	TFILE,
 	TOLD,
 	NALLTYPES,
+
+	TRUNE	= sizeof(Rune)==4? TUINT: TUSHORT,
 };
 enum
 {
@@ -740,7 +742,7 @@ void	gclean(void);
 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*);
diff --git a/sys/src/cmd/cc/cc.y b/sys/src/cmd/cc/cc.y
index 09b788598..eff930b23 100644
--- a/sys/src/cmd/cc/cc.y
+++ b/sys/src/cmd/cc/cc.y
@@ -855,9 +855,9 @@ lstring:
 	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;
@@ -867,16 +867,16 @@ lstring:
 		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:
diff --git a/sys/src/cmd/cc/com.c b/sys/src/cmd/cc/com.c
index 8ff7c4663..a957c3acd 100644
--- a/sys/src/cmd/cc/com.c
+++ b/sys/src/cmd/cc/com.c
@@ -633,10 +633,11 @@ tcomo(Node *n, int f)
 		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));
+				Rune str[1] = {0};
+				outlstring(str, sizeof(Rune));
 				o = outlstring(0, 0);
 			}
 		}
diff --git a/sys/src/cmd/cc/dpchk.c b/sys/src/cmd/cc/dpchk.c
index 99a49ee5a..aa7d468b8 100644
--- a/sys/src/cmd/cc/dpchk.c
+++ b/sys/src/cmd/cc/dpchk.c
@@ -67,13 +67,14 @@ getflag(char *s)
 {
 	Bits flag;
 	int f;
-	char *fmt;
+	char *fmt, *e;
 	Rune c;
 
 	fmt = fmtbuf;
+	e = fmtbuf + sizeof(fmtbuf)-1;
 	flag = zbits;
 	nstar = 0;
-	for(;;) {
+	while(fmt < e){
 		s += chartorune(&c, s);
 		fmt += runetochar(fmt, &c);
 		if(c == 0 || c >= nelem(flagbits))
@@ -175,7 +176,7 @@ pragvararg(void)
 {
 	Sym *s;
 	int n, c;
-	char *t;
+	char *t, *e;
 	Rune r;
 	Type *ty;
 
@@ -225,12 +226,15 @@ cktype:
 	if(c != '"')
 		goto bad;
 	t = fmtbuf;
+	e = t + sizeof(fmtbuf)-1;
 	for(;;) {
 		r = getr();
 		if(r == ' ' || r == '\n')
 			goto bad;
 		if(r == '"')
 			break;
+		if(t >= e)
+			goto bad;
 		t += runetochar(t, &r);
 	}
 	*t = 0;
diff --git a/sys/src/cmd/cc/lex.c b/sys/src/cmd/cc/lex.c
index 68a566ba0..48da7339f 100644
--- a/sys/src/cmd/cc/lex.c
+++ b/sys/src/cmd/cc/lex.c
@@ -467,7 +467,7 @@ l1:
 				yyerror("missing '");
 				peekc = c1;
 			}
-			yylval.vval = convvtox(c, TUSHORT);
+			yylval.vval = convvtox(c, TRUNE);
 			return LUCONST;
 		}
 		if(c == '"') {
@@ -541,15 +541,15 @@ l1:
 			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;
@@ -1027,7 +1027,7 @@ getnsc(void)
 	} else
 		c = GETC();
 	for(;;) {
-		if(!isspace(c))
+		if(c >= Runeself || !isspace(c))
 			return c;
 		if(c == '\n') {
 			lineno++;
diff --git a/sys/src/cmd/cc/pswt.c b/sys/src/cmd/cc/pswt.c
index df1cda4a4..1eb495595 100644
--- a/sys/src/cmd/cc/pswt.c
+++ b/sys/src/cmd/cc/pswt.c
@@ -132,28 +132,28 @@ casf(void)
 }
 
 long
-outlstring(ushort *s, long n)
+outlstring(Rune *s, long n)
 {
-	char buf[2];
-	int c;
+	char buf[sizeof(Rune)];
+	int c, i;
 	long r;
 
 	if(suppress)
 		return nstring;
-	while(nstring & 1)
+	while(nstring % sizeof buf)
 		outstring("", 1);
 	r = nstring;
 	while(n > 0) {
 		c = *s++;
 		if(align(0, types[TCHAR], Aarg1)) {
-			buf[0] = c>>8;
-			buf[1] = c;
+			for(i = sizeof buf; i > 0; c >>= 8)
+				buf[--i] = c;
 		} else {
-			buf[0] = c;
-			buf[1] = c>>8;
+			for(i = 0; i < sizeof buf; c >>= 8)
+				buf[i++] = c;
 		}
-		outstring(buf, 2);
-		n -= sizeof(ushort);
+		outstring(buf, sizeof buf);
+		n -= sizeof buf;
 	}
 	return r;
 }