Drop Python2 support as part of 2.x release?

[jeffwidman]

I realize that many production environments have not migrated to python 3, and normally I would prefer to be the laggard here. However, the timing seems to line up really nicely for dropping python 2 support in the first 2.x release since it coincides with both our release cycle and the deprecation timeline of python 2:

• The official death date of python 2 is Dec 31. Many popular open source libs in the python community are already dropping python2 support left, right, and center: https://python3statement.org
• Given our historical rate of releases, it would not be tremendously out of character for us to wait to cut the next release until Jan 1, 2020. And I would not be opposed to doing it earlier... there's nothing magical about Jan 1, 2020.
• We already need to bump up to 2.0 since I merged #1196.

Plus, we aren't tying anyone's hands... if they still need a python 2 compatible version they can continue to use 1.4.7... it seems to be relatively stable after many, many iterations in the 1.4.x series.

[dpkp]

I don't feel like maintaining py2 + py3 compatibility has been much of a burden for this library. I would prefer to keep compatibility.

[wbarnha]

GitHub runners no longer support Python 2.7, so we can't even test for 2.7 support anymore. I think it's time we do one final release and someone manually tests for 2.7 support, then set the minimum version to 3.8.