[Python-Dev] Problems with unicode_literals

["Martin v. Löwis"]

> That seems a bit *too* strict to me, as long as the Unicode strings
> contain just ASCII. I'm fine with fixing both cases Barry mentioned,
> especially if it otherwise breaks "from __future__ import
> unicode_literals". I expect though that as one tries more things one
> will find more things broken with that mode.

Of course, the proposed patch would widen it to arbitrary Unicode
command options; nothing in the patch restricts it to pure ASCII.

Even when only ASCII characters are used in the option name, we
might still get encoding exceptions or warnings if a non-ASCII byte
string (e.g. from the command line) happens to be compared with the
option name (although I just now couldn't produce such a case).

Regards,
Martin

P.S. optparse already defines a function isbasestring; it might
be better to use that one instead.