Escaping non-printable UTF-8 characters

likema · Apr 6, 2022 · af60fb1 · af60fb1
1 parent d6b1a3b
commit af60fb1
Showing 1 changed file with 53 additions and 46 deletions.
diff --git a/src/yajl_encode.c b/src/yajl_encode.c
@@ -64,64 +64,71 @@ yajl_string_encode(const yajl_print_t print,
             case '\b': escaped = "\\b"; escapedLen = 2; break;
             case '\t': escaped = "\\t"; escapedLen = 2; break;
             default:
-                switch (c & 0xF8) { /* 11111 000 */
-                    case 0: break;
-                    case 0xC0: /* 110 00 000 */
-                    case 0xC8: /* 110 01 000 */
-                    case 0xD0: /* 110 10 000 */
-                    case 0xD8: /* 110 11 000 */
-                        escaped = hexBuf;
-                        utf16 = (c & 0x1F) << 6;
-                        if ((c = str[++end]) & 0x80) {
-                            utf16 |= c & 0x3F;
-                        } else {
-                            utf16 = str[--end];
-                        }
-                        Utf16ToHex (utf16, hexBuf + 2);
-                        break;
-                    case 0xE0: /* 1110 0 000 */
-                    case 0xE8: /* 1110 1 000 */
-                        escaped = hexBuf;
-                        utf16 = (c & 0x0F) << 12;
-                        if ((c = str[++end]) & 0x80) {
-                            utf16 |= (c & 0x3F) << 6;
-
+                if (c < 32) {
+                    escaped = hexBuf;
+                    Utf16ToHex((unsigned short)(c & 0x7F), hexBuf + 2);
+                } else {
+                    switch (c & 0xF8) { /* 11111 000 */
+                        case 0: break;
+                        case 0xC0: /* 110 00 000 */
+                        case 0xC8: /* 110 01 000 */
+                        case 0xD0: /* 110 10 000 */
+                        case 0xD8: /* 110 11 000 */
+                            escaped = hexBuf;
+                            utf16 = (c & 0x1F) << 6;
                             if ((c = str[++end]) & 0x80) {
                                 utf16 |= c & 0x3F;
-                                Utf16ToHex (utf16, hexBuf + 2);
-                                break;
+                            } else {
+                                // Fall through for two byte overlong UTF-8 character.
+                                utf16 = str[--end];
                             }
-                        }
-
-                        Utf16ToHex (str[mark], hexBuf + 2);
-                        end = mark;
-                        break;
-                    case 0xF0: /* 11110 000 */
-                        escaped = hexBuf;
-                        utf16 = (0xD8 << 8) | ((c & 0x07) << 8);
-                        if ((c = str[++end]) & 0x80) {
-                            utf16 |= c << 2;
-
+                            Utf16ToHex (utf16, hexBuf + 2);
+                            break;
+                        case 0xE0: /* 1110 0 000 */
+                        case 0xE8: /* 1110 1 000 */
+                            escaped = hexBuf;
+                            utf16 = (c & 0x0F) << 12;
                             if ((c = str[++end]) & 0x80) {
-                                utf16 |= (c & 0x30) >> 4;
-                                utf16 -= 1 << 6;
-                                Utf16ToHex (utf16, hexBuf + 2);
-
-                                utf16 |= (0xD9 << 8) | (c & 0x0F) << 6;
+                                utf16 |= (c & 0x3F) << 6;
 
                                 if ((c = str[++end]) & 0x80) {
                                     utf16 |= c & 0x3F;
-                                    Utf16ToHex (utf16, hexBuf + 8);
-                                    escapedLen = 12;
+                                    Utf16ToHex (utf16, hexBuf + 2);
                                     break;
                                 }
                             }
-                        }
 
-                        Utf16ToHex (str[mark], hexBuf + 2);
-                        end = mark;
-                        break;
+                            Utf16ToHex (str[mark], hexBuf + 2);
+                            end = mark;
+                            break;
+                        case 0xF0: /* 11110 000 */
+                            escaped = hexBuf;
+                            utf16 = (0xD8 << 8) | ((c & 0x07) << 8);
+                            if ((c = str[++end]) & 0x80) {
+                                utf16 |= c << 2;
+
+                                if ((c = str[++end]) & 0x80) {
+                                    utf16 |= (c & 0x30) >> 4;
+                                    utf16 -= 1 << 6;
+                                    Utf16ToHex (utf16, hexBuf + 2);
+
+                                    utf16 |= (0xD9 << 8) | (c & 0x0F) << 6;
+
+                                    if ((c = str[++end]) & 0x80) {
+                                        utf16 |= c & 0x3F;
+                                        Utf16ToHex (utf16, hexBuf + 8);
+                                        escapedLen = 12;
+                                        break;
+                                    }
+                                }
+                            }
+
+                            Utf16ToHex (str[mark], hexBuf + 2);
+                            end = mark;
+                            break;
+                    }
                 }
+
                 break;
         }
         if (escaped != NULL) {