When do we want to drop Python 2 support for Xarray. For reference, Pandas has a stated drop date for Python 2 of the end of 2018 (this year) and Numpy is slightly later and includes an incremental depreciation, final on Jan. 1, 2020.
We may also consider signing this pledge to help make it clear when/why we're dropping Python 2 support: http://www.python3statement.org/
xref: https://github.com/pandas-dev/pandas/issues/18894, https://github.com/numpy/numpy/pull/10006, https://github.com/python3statement/python3statement.github.io/issues/11
My inclination would be to follow the lead of pandas here, since it's unlikely to make sense to update xarray without its dependencies. So I guess end of 2018 makes sense?
I've just put up a PR on the python3statement repo. I've tentatively also listed 2019-01-01 as a possible target for Xarray 1.0, coincident with dropping Python2, but am now thinking we may need a larger discussion before that comes to any certainty. Anyhow, please take a look at https://github.com/python3statement/python3statement.github.io/pull/102.
My inclination would be to save 1.0 until xarray's data model is stable, which probably means after https://github.com/pydata/xarray/issues/1603. If the timing works out for dropping Python 2 support that would be great, but I wouldn't want to force it.
@pydata/xarray - any objections to the timeline here? We will need to close this before https://github.com/python3statement/python3statement.github.io/pull/102 can be merged.
I don't object to the timeline, but I don't think we're ready to commit to supporting Python 2 for all 0.X releases and stopping it with 1.0. The particular version number feels premature to me.
Great. No argument there. I'll adjust the PR on the python3statement to reflect that.
I fully agree with the proposed changes!
Most helpful comment
My inclination would be to follow the lead of pandas here, since it's unlikely to make sense to update xarray without its dependencies. So I guess end of 2018 makes sense?