[5.5.3]: Fix GRIB IOSP time coords from intervals, for 2-D time datasets

[ethanrd]

Description of Changes

Fix generation of time coordinate variables from time intervals for 2-D time datasets. Ensure resulting time coordinate variables are strictly monotonically increasing. Handle when time intervals are irregular and overlapping.

See issue #1060.

PR Checklist

• Link to any issues that the PR addresses
• Add labels
• Open as a draft PR
  until ready for review
• Make sure GitHub tests pass
• Mark PR as "Ready for Review"

[tdrwenski]

I ran all the tests on Jenkins against your branch and one is failing: https://jenkins-aws.unidata.ucar.edu/view/Users/job/tara-netcdf-java/35/

In the TestGribCompareBuilders tests it looks like it is comparing the builder/ non-builder versions for all files in ../grib/src/test/data/ and it looks like there's a difference for the new file you added. I do not believe this is actually related to your changes (I ran the test with that file on maint-5.x and it also failed). Since it's unrelated and I'm not sure how important it is that the with/without builders return identical results, maybe it's easiest to just ignore that new test file in that test and possibly fix it later.

[ethanrd]

Agree, probably easier to ignore the new test file. Are there other test files excluded from this test?

[tdrwenski]

Agree, probably easier to ignore the new test file. Are there other test files excluded from this test?

Just pushed a fix to ignore that test case!