Comment 2 for bug 563692

Problem appears to be with any optparse.Option that specifies type=str, which will naturally raise on non-ascii values. Either a type function that reraises a useful error should be substituted, or a higher layer should catch the UnicodeError and handle it there.