Serhiy Storchaka <storchaka+cpyt...@gmail.com> added the comment:
Stefan's arguments looked reasonable to me. I supported deprecating Element.__bool__. This was an old plan, although we failed with deprecation in the C implementation (this is not the first oversight in deprecations in ElementTree). But there is a problem: removing __bool__ will not change the value of bool(element). It will still return False for an Element without children, because bool() falls back to __len__ if __bool__ is not defined, and Element.__len__ is defined. So we will need either to define Element.__bool__ always returning True, and this will be the first case when __bool__ and __len__ return inconsistent values, or to deprecate __len__ (and perhaps __getitem__) that goes too far from the initial plan. So now I am not sure. Making Element.__bool__ always returning True or raising an exception could help to fix some bugs in user code. But this differs from common behavior of sequence-like types in Python. On other hand, we have a precedence: bool(NotImplemented) raises an exception now (this also can save from common errors). When I started to write this message I was -0 for changing Element.__bool__. Now I am +0, after considering all arguments that came in my mind, but still have doubts. Can we ask some of the language ideologues? ---------- nosy: +gvanrossum _______________________________________ Python tracker <rep...@bugs.python.org> <https://bugs.python.org/issue38941> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
