In a few days, we will be celebrating the 5th anniversary from the release of python3.6. With python3.10 released a few months ago, do you believe that maintainers of python packages should still provide support for python3.6? This of course comes with plenty of trade-offs, such us not being able to update dependencies, additional code and testing to handle python version incompatibilities and the list goes on. As a maintainer of a package that was created after python3.6, I have got multiple requests to provide support for it, which is of course something that I would be keen to avoid. The answer depends on the library domain I guess, but I would like to get a holistic view of the matter.
Python 3.6 EOL date is 2021-12-23 (in 2 weeks) [0]. There's also PyPI download stats you can check for your package. E.g. requests has ~20% downloads from 3.6 [1], but for your package it might be way lower, so you can base your decision on that.<p>[0] <a href="https://devguide.python.org/#status-of-python-branches" rel="nofollow">https://devguide.python.org/#status-of-python-branches</a><p>[1] <a href="https://pypistats.org/packages/requests" rel="nofollow">https://pypistats.org/packages/requests</a> (stats for my packages look weird starting November btw, so might need to look at October data)