Éric Araujo <mer...@netwok.org> added the comment: self.distribution.username: >>>> I’d prefer a clearer comparison, please use “is [not] None” and parens. >>> Are you sure you want an empty username in config file to override name set from >>> 'register' command? >> This is a crazy case. PyPI won’t allow it to pass, so I’m okay with it. > That's why I don't want to make distinction between "None" username and empty > username. They are both the same in this context.
That’s not what I meant. IMO, None means not provided and an empty string (i.e. an empty value in the config file) is a bug introduced by the user. not >> “password”. > That's magically patched in setUp method. I don’t understand. The RawInput monkey-patch will give '1', 'tarek' and 'n', which are anwsers to “choice”, “username” and “save password”, but the code should ask for “choice”, “username”, “password”, “save password”. http://codereview.appspot.com/2874041/diff/13001/command/upload.py File command/upload.py (right): http://codereview.appspot.com/2874041/diff/13001/command/upload.py#newcode53 command/upload.py:53: if not self.username and self.distribution.username: A bugfix should not change the API like this. Either make it a private name like _username, or (better) use dist.get_command_obj('register').username http://codereview.appspot.com/2874041/ ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue10361> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com