> The function first normalizes the "@" away and then looks for it. Is that > the expected behaviour?
I believe this functionality is broken by design. Python can't possibly know correctly what each locale name on each system means, and what encoding is used in the locale. Instead, the system's API to find out the encoding should be used, as exposed in locale.getpreferredencoding(). Regards, Martin -- http://mail.python.org/mailman/listinfo/python-list