gc: disallow NUL byte, catch more invalid UTF-8, test

R=ken2, ken3
CC=golang-dev
https://golang.org/cl/209041

src/cmd/gc/lex.c

@@ -520,18 +520,19 @@ l0:
 		ncp = 8;
 
 		for(;;) {
-			if(clen == ncp) {
-				cp = remal(cp, clen, ncp);
+			if(clen+UTFmax > ncp) {
+				cp = remal(cp, ncp, ncp);
 				ncp += ncp;
 			}
-			c = getc();
+			c = getr();
 			if(c == EOF) {
 				yyerror("eof in string");
 				break;
 			}
 			if(c == '`')
 				break;
-			cp[clen++] = c;
+			rune = c;
+			clen += runetochar(cp+clen, &rune);
 		}
 
 	strlit:
@@ -821,28 +822,16 @@ talph:
 	 */
 	for(;;) {
 		if(c >= Runeself) {
-			for(c1=0;;) {
-				cp[c1++] = c;
-				if(fullrune(cp, c1)) {
-					chartorune(&rune, cp);
-					if(isfrog(rune)) {
-						yyerror("illegal character 0x%ux", rune);
-						goto l0;
-					}
-					// 0xb7 · is used for internal names
-					if(!isalpharune(rune) && !isdigitrune(rune) && rune != 0xb7)
-						yyerror("invalid identifier character 0x%ux", rune);
-					break;
-				}
-				c = getc();
-			}
-			cp += c1;
-			c = getc();
-			continue;
-		}
-		if(!isalnum(c) && c != '_')
+			ungetc(c);
+			rune = getr();
+			// 0xb7 · is used for internal names
+			if(!isalpharune(rune) && !isdigitrune(rune) && rune != 0xb7)
+				yyerror("invalid identifier character 0x%ux", rune);
+			cp += runetochar(cp, &rune);
+		} else if(!isalnum(c) && c != '_')
 			break;
-		*cp++ = c;
+		else
+			*cp++ = c;
 		c = getc();
 	}
 	*cp = 0;
@@ -1054,8 +1043,10 @@ getc(void)
 
 	switch(c) {
 	case 0:
-		if(curio.bin != nil)
+		if(curio.bin != nil) {
+			yyerror("illegal NUL byte");
 			break;
+		}
 	case EOF:
 		return EOF;
 
@@ -1097,10 +1088,11 @@ loop:
 	c = chartorune(&rune, str);
 	if(rune == Runeerror && c == 1) {
 		lineno = lexlineno;
-		yyerror("illegal UTF-8 sequence in comment or string");
+		yyerror("illegal UTF-8 sequence");
 		flusherrors();
+		print("\t");
 		for(c=0; c<i; c++)
-			print(" %.2x", *(uchar*)(str+c));
+			print("%s%.2x", c > 0 ? " " : "", *(uchar*)(str+c));
 		print("\n");
 	}
 	return rune;
@@ -1209,11 +1201,11 @@ oct:
 			l = l*8 + c-'0';
 			continue;
 		}
-		yyerror("non-oct character in escape sequence: %c", c);
+		yyerror("non-octal character in escape sequence: %c", c);
 		ungetc(c);
 	}
 	if(l > 255)
-		yyerror("oct escape value > 255: %d", l);
+		yyerror("octal escape value > 255: %d", l);
 
 	*val = l;
 	return 0;

src/cmd/gc/subr.c

@@ -1525,6 +1525,7 @@ Zconv(Fmt *fp)
 	Rune r;
 	Strlit *sp;
 	char *s, *se;
+	int n;
 
 	sp = va_arg(fp->args, Strlit*);
 	if(sp == nil)
@@ -1533,8 +1534,15 @@ Zconv(Fmt *fp)
 	s = sp->s;
 	se = s + sp->len;
 	while(s < se) {
-		s += chartorune(&r, s);
+		n = chartorune(&r, s);
+		s += n;
 		switch(r) {
+		case Runeerror:
+			if(n == 1) {
+				fmtprint(fp, "\\x%02x", *(s-1));
+				break;
+			}
+			// fall through
 		default:
 			if(r < ' ') {
 				fmtprint(fp, "\\x%02x", r);

test/nul.go

@@ -0,0 +1,58 @@
+// $G $D/$F.go && $L $F.$A && ./$A.out >tmp.go &&
+// errchk $G -e tmp.go
+// rm -f tmp.go
+
+// Copyright 2009 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Test source files and strings containing NUL and invalid UTF-8.
+
+package main
+
+import (
+	"fmt"
+	"os"
+)
+
+func main() {
+	var s = "\xc2\xff"
+	var t = "\xd0\xfe"
+	var u = "\xab\x00\xfc"
+
+	if len(s) != 2 || s[0] != 0xc2 || s[1] != 0xff ||
+		len(t) != 2 || t[0] != 0xd0 || t[1] != 0xfe ||
+		len(u) != 3 || u[0] != 0xab || u[1] != 0x00 || u[2] != 0xfc {
+		println("BUG: non-UTF-8 string mangled");
+		os.Exit(2)
+	}
+
+	fmt.Print(`
+package main
+
+var x = "in string ` + "\x00" + `"	// ERROR "NUL"
+
+var y = ` + "`in raw string \x00 foo`" + `  // ERROR "NUL"
+
+// in comment ` + "\x00" + `  // ERROR "NUL"
+
+/* in other comment ` + "\x00" + ` */ // ERROR "NUL"
+
+/* in source code */ ` + "\x00" + `// ERROR "NUL"
+
+var xx = "in string ` + "\xc2\xff" + `" // ERROR "UTF-8"
+
+var yy = ` + "`in raw string \xff foo`" + `  // ERROR "UTF-8"
+
+// in comment ` + "\xe2\x80\x01" + `  // ERROR "UTF-8"
+
+/* in other comment ` + "\xe0\x00\x00" + ` */ // ERROR "UTF-8"
+
+/* in variable name */
+var z` + "\xc1\x81" + ` int // ERROR "UTF-8"
+
+/* in source code */ ` + "\xc2A" + `// ERROR "UTF-8"
+
+`)
+}
+