473         # surrogate-encode any characters that don't round-trip

   473         pass

   474         s2 = s.decode('utf-8', 'ignore').encode('utf-8')

   474 

   475         r = ""

   475     r = ""

   476         pos = 0

   476     pos = 0

   477         for c in s:

   477     l = len(s)

   478             if s2[pos:pos + 1] == c:

   478     while pos < l:

   479                 r += c

   479         try:

   480                 pos += 1

   480             c = getutf8char(s, pos)

   481             else:

   481             pos += len(c)

   482                 r += unichr(0xdc00 + ord(c)).encode('utf-8')

   482         except UnicodeDecodeError:

   483         return r

   483             c = unichr(0xdc00 + ord(s[pos])).encode('utf-8')

   484             pos += 1

   485         r += c

   486     return r