01-JSON_parse.patch

diff --git a/ext/json/JSON_parser.c b/ext/json/JSON_parser.c
index dd832a7..3749998 100644
--- a/ext/json/JSON_parser.c
+++ b/ext/json/JSON_parser.c
@@ -353,40 +353,59 @@ use_string:
 
 static void utf16_to_utf8(smart_str *buf, unsigned short utf16)
 {
-    if (utf16 < 0x80)
-    {
-        smart_str_appendc(buf, (unsigned char) utf16);
-    }
-    else if (utf16 < 0x800)
-    {
-        smart_str_appendc(buf, 0xc0 | (utf16 >> 6));
-        smart_str_appendc(buf, 0x80 | (utf16 & 0x3f));
-    }
-    else if ((utf16 & 0xfc00) == 0xdc00
-                && buf->len >= 3
-                && ((unsigned char) buf->c[buf->len - 3]) == 0xed
-                && ((unsigned char) buf->c[buf->len - 2] & 0xf0) == 0xa0
-                && ((unsigned char) buf->c[buf->len - 1] & 0xc0) == 0x80)
-    {
-        /* found surrogate pair */
-        unsigned long utf32;
-
-        utf32 = (((buf->c[buf->len - 2] & 0xf) << 16)
-                    | ((buf->c[buf->len - 1] & 0x3f) << 10)
-                    | (utf16 & 0x3ff)) + 0x10000;
-        buf->len -= 3;
-
-        smart_str_appendc(buf, (unsigned char) (0xf0 | (utf32 >> 18)));
-        smart_str_appendc(buf, 0x80 | ((utf32 >> 12) & 0x3f));
-        smart_str_appendc(buf, 0x80 | ((utf32 >> 6) & 0x3f));
-        smart_str_appendc(buf, 0x80 | (utf32 & 0x3f));
-    }
-    else
-    {
-        smart_str_appendc(buf, 0xe0 | (utf16 >> 12));
-        smart_str_appendc(buf, 0x80 | ((utf16 >> 6) & 0x3f));
-        smart_str_appendc(buf, 0x80 | (utf16 & 0x3f));
+    if ((utf16 & 0xfc00) == 0xdc00) {
+
+        if (buf->len >= 3
+            && ((unsigned char) buf->c[buf->len - 3] == 0xed)
+            && ((unsigned char) buf->c[buf->len - 2] & 0xf0) == 0xa0
+            && ((unsigned char) buf->c[buf->len - 1] & 0xc0) == 0x80
+        ) {
+            /* found surrogate pair */
+            unsigned long utf32;
+
+            utf32 = (((buf->c[buf->len - 2] & 0xf) << 16)
+                        | ((buf->c[buf->len - 1] & 0x3f) << 10)
+                        | (utf16 & 0x3ff)) + 0x10000;
+            buf->len -= 3;
+
+            smart_str_appendc(buf, (unsigned char) (0xf0 | (utf32 >> 18)));
+            smart_str_appendc(buf, 0x80 | ((utf32 >> 12) & 0x3f));
+            smart_str_appendc(buf, 0x80 | ((utf32 >> 6) & 0x3f));
+            smart_str_appendc(buf, 0x80 | (utf32 & 0x3f));
+        } else {
+            buf->len -= 3;
+            smart_str_appendl(buf, "\xef\xbf\xbd", 3);
+        }
+
+    } else {
+
+        if (buf->len >= 3
+            && ((unsigned char) buf->c[buf->len - 3] == 0xed)
+            && ((unsigned char) buf->c[buf->len - 2] & 0xf0) == 0xa0
+            && ((unsigned char) buf->c[buf->len - 1] & 0xc0) == 0x80
+        ) {
+            buf->len -= 3;
+            smart_str_appendl(buf, "\xef\xbf\xbd", 3);
+        }
+
+        if (utf16 < 0x80)
+        {
+            smart_str_appendc(buf, (unsigned char) utf16);
+        }
+        else if (utf16 < 0x800)
+        {
+            smart_str_appendc(buf, 0xc0 | (utf16 >> 6));
+            smart_str_appendc(buf, 0x80 | (utf16 & 0x3f));
+        }
+        else
+        {
+            smart_str_appendc(buf, 0xe0 | (utf16 >> 12));
+            smart_str_appendc(buf, 0x80 | ((utf16 >> 6) & 0x3f));
+            smart_str_appendc(buf, 0x80 | (utf16 & 0x3f));
+        }
     }
+
+
 }
 
 static void attach_zval(JSON_parser jp, int up, int cur, smart_str *key, int assoc TSRMLS_DC)

02-test.php

<?php
var_dump(
  "\xef\xbf\xbd" === json_decode('"\udc00"'),
  "\xef\xbf\xbd"."\xed\xa0\x80" === json_decode('"\ud800\ud800"'),
  "\xed\xa0\x80" === json_decode('"\ud800"')
);