There is a known bug that impacts some very rare edge cases I haven't been able to easily identify when building the coregistration tree (I think? or maybe the SBAS dates, something to do with dates... apologies - it's been many months since I worked on this project)
This issue is to just raise this as a known problem for new developers (eg: @NMoghaddam ) so the test failure doesn't come up randomly and confuse them, the test added to sometimes identify the issue is tests/test_date_utils.py::test_simplify_dates_fuzz
somehow one side ends up being None instead of a date, should be a simple fix - it just needs someone to sit down figure out how/where the code could pass None into the function.
Just to clarify, this has never hit a production or even a test run on NCI. We've only seen it in our unit tests so far, so we know it could be an issue (but so far it hasn't impacted anyone).
There is a known bug that impacts some very rare edge cases I haven't been able to easily identify when building the coregistration tree (I think? or maybe the SBAS dates, something to do with dates... apologies - it's been many months since I worked on this project)
Wwe created a fuzz test to try and sometimes hit this issue, for an example of the failure see: https://github.com/GeoscienceAustralia/PyGamma/actions/runs/3295030571/jobs/5433157439
This issue is to just raise this as a known problem for new developers (eg: @NMoghaddam ) so the test failure doesn't come up randomly and confuse them, the test added to sometimes identify the issue is
tests/test_date_utils.py::test_simplify_dates_fuzz
somehow one side ends up being
None
instead of a date, should be a simple fix - it just needs someone to sit down figure out how/where the code could pass None into the function.