In light of the official EOL date for Python 2 at the end of this year, and the fact that most of our critical python dependencies seem to be on board with this timeline, after having had a lazy decade to make the transition, it'll be great to finally pull the plug on supporting python 2.7. We would:
drop 2.7 from the Travis test matrix
consolidate the requirements-*.txt files to assume py3 only
remove the conditionals that decorate our unit/integration tests that allow py2 to skip them (e.g. all the snakemake tests would become mandatory)
anything else I forgot
I don't see a particular reason to wait to Dec 31 (or wait at all), do you?
In light of the official EOL date for Python 2 at the end of this year, and the fact that most of our critical python dependencies seem to be on board with this timeline, after having had a lazy decade to make the transition, it'll be great to finally pull the plug on supporting python 2.7. We would:
requirements-*.txtfiles to assume py3 onlyI don't see a particular reason to wait to Dec 31 (or wait at all), do you?