> Posted https://mail.python.org/archives/list/python-dev@python.org/thread/4UZJYAZL3NHRAGN5WAMJC4IHAHEXF3QF/ to see if anyone else wants to weigh in.

Mark Shannon listed flaws in the PEP 532 and suggest to withdraw this PEP. Honestly, I'm open to any solution. But this issue is a concrete regression of Python 3.8 with a concrete use case, whereas Mark only lists theoretical enhancements.

I would like to fix the regression first. There are users of the PEP 532, we cannot simply ignore them, withdraw the PEP and remove the feature immediately without warning users.

If someone wants to withdraw the PEP, it has to go through the regular deprecation process with a slow transition. Last time we pushed too many incompatible changes (Python 2 to Python 3 transition), it was a mess. Moreover, I would prefer to have at least one Python release (if not two or more) which provides two options, deprecated way and a new better way, to have a smooth transition.