From f68498a6fdb421483d9aebb45527452f6255bb68 Mon Sep 17 00:00:00 2001
From: Luke Shumaker <lukeshu@lukeshu.com>
Date: Sat, 25 Feb 2023 16:17:01 -0700
Subject: jsonparse: Don't show raw bytes as Unicode

---
 reencode.go | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'reencode.go')

diff --git a/reencode.go b/reencode.go
index 7439bf0..8b08aad 100644
--- a/reencode.go
+++ b/reencode.go
@@ -352,7 +352,7 @@ func (enc *ReEncoder) Close() error {
 
 // isRune=false indicates that 'c' is a raw byte from invalid UTF-8.
 func (enc *ReEncoder) handleRune(c rune, size int, isRune bool) {
-	t, err := enc.par.HandleRune(c)
+	t, err := enc.par.HandleRune(c, isRune)
 	if err != nil {
 		enc.err = &ReEncodeSyntaxError{
 			Err:    err,
-- 
cgit v1.2.3-2-g168b


From 7bd0072b5896bfc4172b6bda778cf149dd6282fa Mon Sep 17 00:00:00 2001
From: Luke Shumaker <lukeshu@lukeshu.com>
Date: Sat, 25 Feb 2023 16:17:06 -0700
Subject: reencode: Fix the byte count for partial writes

---
 reencode.go | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

(limited to 'reencode.go')

diff --git a/reencode.go b/reencode.go
index 8b08aad..fd36875 100644
--- a/reencode.go
+++ b/reencode.go
@@ -243,10 +243,13 @@ func (enc *ReEncoder) getRuneFromString(str string, pos int) (c rune, size int,
 // but *ReEncoder does because it transforms the data written to it,
 // and the number of bytes written may be wildly different than the
 // number of bytes handled.
+//
+//nolint:dupl // Yes, this is mostly a duplicate of .WriteString().
 func (enc *ReEncoder) Write(str []byte) (int, error) {
 	if len(str) == 0 {
 		return 0, nil
 	}
+	origBufLen := enc.bufLen
 	var n int
 	for {
 		c, size, full, isRune := enc.getRuneFromBytes(str, n)
@@ -261,14 +264,14 @@ func (enc *ReEncoder) Write(str []byte) (int, error) {
 			return len(str), nil
 		}
 		if enc.utf == InvalidUTF8Error && !isRune {
-			return n, &ReEncodeSyntaxError{
+			return n - origBufLen, &ReEncodeSyntaxError{
 				Offset: enc.inputPos,
 				Err:    fmt.Errorf("invalid UTF-8: %#02x", c),
 			}
 		}
 		enc.handleRune(c, size, isRune)
 		if enc.err != nil {
-			return n, enc.err
+			return n - origBufLen, enc.err
 		}
 		n += size
 	}
@@ -276,10 +279,13 @@ func (enc *ReEncoder) Write(str []byte) (int, error) {
 
 // WriteString implements io.StringWriter; it does what you'd expect,
 // but see the notes on the Write method.
+//
+//nolint:dupl // Yes, this is mostly a duplicate of .Write().
 func (enc *ReEncoder) WriteString(str string) (int, error) {
 	if len(str) == 0 {
 		return 0, nil
 	}
+	origBufLen := enc.bufLen
 	var n int
 	for {
 		c, size, full, isRune := enc.getRuneFromString(str, n)
@@ -294,14 +300,14 @@ func (enc *ReEncoder) WriteString(str string) (int, error) {
 			return len(str), nil
 		}
 		if enc.utf == InvalidUTF8Error && !isRune {
-			return n, &ReEncodeSyntaxError{
+			return n - origBufLen, &ReEncodeSyntaxError{
 				Offset: enc.inputPos,
 				Err:    fmt.Errorf("invalid UTF-8: %#02x", c),
 			}
 		}
 		enc.handleRune(c, size, isRune)
 		if enc.err != nil {
-			return n, enc.err
+			return n - origBufLen, enc.err
 		}
 		n += size
 	}
-- 
cgit v1.2.3-2-g168b


From 22edcf6a68a057ed04368d5f78c8ba3ddfee8d57 Mon Sep 17 00:00:00 2001
From: Luke Shumaker <lukeshu@lukeshu.com>
Date: Sat, 25 Feb 2023 11:11:36 -0700
Subject: reencode: Improve the error messages for trailing partial-UTF-8

---
 reencode.go | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)

(limited to 'reencode.go')

diff --git a/reencode.go b/reencode.go
index fd36875..c19e296 100644
--- a/reencode.go
+++ b/reencode.go
@@ -329,9 +329,21 @@ func (enc *ReEncoder) WriteRune(c rune) (n int, err error) {
 // if enc.AllowMultipleValues is set.
 func (enc *ReEncoder) Close() error {
 	if enc.bufLen > 0 {
-		return &ReEncodeSyntaxError{
-			Offset: enc.inputPos,
-			Err:    fmt.Errorf("%w: unflushed unicode garbage: %q", io.ErrUnexpectedEOF, enc.buf[:enc.bufLen]),
+		if enc.utf == InvalidUTF8Error {
+			return &ReEncodeSyntaxError{
+				Offset: enc.inputPos,
+				Err:    fmt.Errorf("truncated UTF-8: %q", enc.buf[:enc.bufLen]),
+			}
+		}
+		for i := 0; i < enc.bufLen; i++ {
+			if enc.utf == InvalidUTF8Replace {
+				enc.handleRune(utf8.RuneError, 1, true)
+			} else {
+				enc.handleRune(rune(enc.buf[i]), 1, false)
+			}
+			if enc.err != nil {
+				return enc.err
+			}
 		}
 	}
 	if _, err := enc.par.HandleEOF(); err != nil {
-- 
cgit v1.2.3-2-g168b