"Juergen Hermann" <jh@web.de> writes: > Explicit: stream.write(u.encode('utf-8')) > > Implicit: Use codecs.lookup('utf-8') to get reader/writer wrappers, then > you say utf8_stdout = utf8_wrapper(sys.stdout) Since Python 2.2, it is better to use codecs.getwriter("utf-8"), since that saves the trouble of knowing in which order the result of .lookup is. Regards, Martin