R. David Murray added the comment: I think a fully logical fix could be implemented for 3.4 (after I commit this), because as you say it is *unlikely* that anyone is relying on the behaviors mentioned as "backward compatibility issues" in the added tests. However, it is *possible* someone is dealing with a quirky situation where some other broken software is using these flags in key/value form, and depending on it. Not likely, no (especially since the value would be lost on re-serialization), but since we *can* fix the issue without risking breaking such code, we should, for the maintenance release.
---------- _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue16611> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com