Remove Python <=3.6 support

[pelson]

I'm not sure of the schedule, but NEP29 gives some indication that Python 3.6 is becoming legacy (and no longer receives security support patches as per https://endoflife.date/python). At the same time, the CI pipelines now only test Python 3.7+ (upto Python 3.10) - a recent update reasonably (IMO) chose not to put significant effort into testing Python 3.6 in CI, therefore there is no guarantee that the next version of JPype will actually support it.

Given all of the above, I went through the codebase to try to find places where we can remove old bits of code, and update the minimum dependency of JPype to Python 3.7. I am comfortable if this PR is rejected or put on hold (so long as there is a condition for un-holding it put in place).

[codecov]

Codecov Report

Merging #1061 (b93eb28) into master (61c98a1) will decrease coverage by 0.00%.
The diff coverage is 50.00%.

@@            Coverage Diff             @@
##           master    #1061      +/-   ##
==========================================
- Coverage   88.65%   88.65%   -0.01%     
==========================================
  Files         112      112              
  Lines       10236    10232       -4     
  Branches     4012     4012              
==========================================
- Hits         9075     9071       -4     
  Misses        702      702              
  Partials      459      459              

Impacted Files	Coverage Δ
jpype/_core.py	95.39% <ø> (-0.09%)	⬇️
jpype/_pykeywords.py	100.00% <ø> (ø)
jpype/imports.py	78.78% <ø> (ø)
jpype/pickle.py	88.09% <ø> (ø)
jpype/protocol.py	98.71% <ø> (ø)
native/python/pyjp_class.cpp	85.73% <ø> (ø)
native/python/pyjp_module.cpp	83.24% <ø> (ø)
native/python/pyjp_array.cpp	78.63% <50.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 61c98a1...b93eb28. Read the comment docs.

[Thrameos]

LGTM. Excellent work.

[Thrameos]

I think we should include this in the next release. Though the decision should go to @marscher .

Thanks for taking care of this. I am unfortunately pulling a "no lifer" on a work project creating a 30k line code from scratch to act as a realtime spectral gamma server for an augmented reality trainer. (>65 hr/wk) The good news is that I beat on jpype paths hourly (largely without hitting anything that requires patching, though the inability to pass single abstract methods (SAM) to lmfit/scipy optimize) is certainly at the annoyance level).

[pelson]

Thanks for the feedback and approval. I found a few more PY_VERSION_HEX cases, so added a commit.

[ap--]

Hi everyone,

Together with this PR it might be worth running pyupgrade --py37-plus over the python part of the codebase to upgrade older style syntax in an automated way.

Cheers,
Andreas 😃

[Thrameos]

Minor issue. The release scripts got missed. These are vital as they decide which artifacts will get built. Thus adding or removing support has to update those hidden scripts in .azure