Drop Python 3.8 support

[djhoese]

It's what all the cool kids are doing.

[djhoese]

Note I also removed some old multiprocessing import code in setup.py that said it could be removed when Python 2 support was dropped.

[codecov]

Codecov Report

Merging #499 (0cc31d4) into main (b8ecc71) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main     #499   +/-   ##
=======================================
  Coverage   94.33%   94.33%           
=======================================
  Files          74       74           
  Lines       12947    12947           
=======================================
  Hits        12214    12214           
  Misses        733      733           

Flag	Coverage Δ
unittests	94.33% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[pnuu]

Gnaah. Don't know what I did there. But LGTM.

One thought I had was that maybe the python_requires shouldn't be so strict, most of the code will work fine with earlier versions. Though even in that case we should state that only the tested Python versions are actually supported.

[djhoese]

I'm noticing codecov is complaining of lost coverage in the multiprocessing parts (_spatial_mp.py) for modules that haven't been changed, even in #498 which doesn't have these python version changes. Makes me wonder if something has changed in some package (numpy? pyproj?) that is no longer accessing the __iter__ or other parts of these modules.

[djhoese]

Though even in that case we should state that only the tested Python versions are actually supported.

The problem is that python_requires can control what versions of the library PyPI returns. So if we make a release that supports Python 3.7 and then make a bug fix release that introduces syntax that is 3.8+ only then we'd have to remember to drop support in a bug fix version (or release a new minor release just for a bug fix). My point is I think CI needs to stay in sync with what we advertise as supported and it might be better to do it sooner rather than later. 🤷‍♂️

[pnuu]

Ah, right. That'd get complicated and confusing quickly. So yeah, lets stay in sync with the CI.

[coveralls]

Coverage: 93.862%. Remained the same when pulling 0cc31d4 on djhoese:drop-py38 into b8ecc71 on pytroll:main.