py.std.
From the viewpoint of making it easier for new contributors this is another abstraction to be learned. It also trips up IDEs like PyCharm, loosing all the perks of static code analysis and code completion.
Thereofore I would like to replace the remainung (16) occurrences of that with the traditional way of using stdlib modules, via importing them directly.
pytest recently did the same (with the goal of getting rid of pylib entirely at some point), so 馃憤
thanks, good to know - getting rid of it completely in the long run might also be a good idea.
also what about using six as Python 2/3 compatibility? there are some places in the code base where we now naively do these compatibility with if else ?
@gaborbernat good point, but please lets open a new issue for that
I opened an issue that outlines all changes necessary to get rid of py. - which also includes the py2/py3 compat layer, because py.builtin is partially used for that: https://github.com/tox-dev/tox/issues/610
Most helpful comment
@gaborbernat good point, but please lets open a new issue for that