Been focusing on 3.1.1, what we found is that the attached script returns:<br>'\ud801\udc51'<br>'\U00010451'<br><br>This was attached to ensure it transfers properly over the email list :-)<br><br>And, sadly, the workaround is adding .encode('utf-16').decode('utf-16'). It appears that utf-8 support is bugged.<br>
<br>Make sure that you have a "wide" Python build for this, you can test that with:<br>>>> import sys<br>>>> sys.maxunicode<br>1114111<br><br>A narrow build will report 65536.<br>