Handle None being passed explicitly to startJVM

[pelson]

Following-up with my comment in https://github.com/jpype-project/jpype/pull/1062/files#r1182683565.

Apologies to @marscher for missing this in #1062.

[codecov]

Codecov Report

Merging #1134 (507ede7) into master (fffec42) will not change coverage.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master    #1134   +/-   ##
=======================================
  Coverage   87.84%   87.84%           
=======================================
  Files         112      112           
  Lines       10276    10276           
  Branches     4032     4032           
=======================================
  Hits         9027     9027           
  Misses        698      698           
  Partials      551      551           

Impacted Files	Coverage Δ
jpype/_core.py	95.61% <100.00%> (ø)

[marscher]

Thanks, I've overlooked this one too. Do you think there are more spots like this? :)
Since this is a crucial entry point, we should probably create an extensive test matrix covering all combination and expected outcomes. I suggest using pytest.parametrize for this purpose.
Could you do that @pelson?

[pelson]

Do you think there are more spots like this? :)
Since this is a crucial entry point, we should probably create an extensive test matrix covering all combination and expected outcomes. I suggest using pytest.parametrize for this purpose.

Already the test_startup cases are pretty extensive. Naturally, as part of this work I extended those cases, and tried to be fairly systematic about checking things, but there can always be more things that slip through (particularly if undocumented). I'll take another look to see if there are any more obvious cases, but otherwise we can have reasonable confidence, and I'll try to address any issues quickly if they come up.