encoding/json: coerce invalid UTF-8 to valid UTF-8 during Marshal

In practice, rejecting an entire structure due to a single invalid byte
in a string is just too picky, and too hard to track down.
Be consistent with the bulk of the standard library by converting
invalid UTF-8 into UTF-8 with replacement runes.

R=golang-dev, crawshaw
CC=golang-dev
https://golang.org/cl/11211045

doc/go1.2.txt

@@ -17,6 +17,7 @@ crypto/sha1: Sum function to simplify hashing (CL 10571043).
 crypto/sha256: Sum256 and Sum224 functions to simplify hashing (CL 10629043).
 crypto/sha512: Sum512 and Sum384 functions to simplify hashing (CL 10630043).
 crypto/tls: add support for TLS 1.1. (CL 7872043).
+encoding/json: accept but correct invalid UTF-8 in Marshal (CL 11211045).
 flag: add Getter interface (CL 10472043).
 fmt: indexed access to arguments in Printf etc. (CL 9680043).
 go/build: support including C++ code with cgo (CL 8248043).

src/pkg/encoding/json/decode_test.go

@@ -393,15 +393,10 @@ func TestMarshal(t *testing.T) {
 
 func TestMarshalBadUTF8(t *testing.T) {
 	s := "hello\xffworld"
+	const enc = `"hello\ufffdworld"`
 	b, err := Marshal(s)
-	if err == nil {
-		t.Fatal("Marshal bad UTF8: no error")
-	}
-	if len(b) != 0 {
-		t.Fatal("Marshal returned data")
-	}
-	if _, ok := err.(*InvalidUTF8Error); !ok {
-		t.Fatalf("Marshal did not return InvalidUTF8Error: %T %v", err, err)
+	if string(b) != enc || err != nil {
+		t.Errorf("Marshal(%q) = %#q, %v, want %#q, nil", s, b, err, enc)
 	}
 }
 

src/pkg/encoding/json/encode.go

@@ -209,8 +209,12 @@ func (e *UnsupportedValueError) Error() string {
 	return "json: unsupported value: " + e.Str
 }
 
-// An InvalidUTF8Error is returned by Marshal when attempting
-// to encode a string value with invalid UTF-8 sequences.
+// Before Go 1.2, an InvalidUTF8Error was returned by Marshal when
+// attempting to encode a string value with invalid UTF-8 sequences.
+// As of Go 1.2, Marshal instead coerces the string to valid UTF-8 by
+// replacing invalid bytes with the Unicode replacement rune U+FFFD.
+// This error is no longer generated but is kept for backwards compatibility
+// with programs that might mention it.
 type InvalidUTF8Error struct {
 	S string // the whole string value that caused the error
 }
@@ -555,7 +559,13 @@ func (e *encodeState) string(s string) (int, error) {
 		}
 		c, size := utf8.DecodeRuneInString(s[i:])
 		if c == utf8.RuneError && size == 1 {
-			e.error(&InvalidUTF8Error{s})
+			if start < i {
+				e.WriteString(s[start:i])
+			}
+			e.WriteString(`\ufffd`)
+			i += size
+			start = i
+			continue
 		}
 		// U+2028 is LINE SEPARATOR.
 		// U+2029 is PARAGRAPH SEPARATOR.