While working on #23, I have run into some challenges because of our messy testing setup. The property-based testing approach seems to work well but there are a few things that need to be improved:
[ ] Make adding a new property to a test significantly simpler, with less boilerplate.
[ ] Run tests to completion even if one property fails.
[ ] Record more granular test results.
[ ] Output DEMs and other information from failing tests.
[x] #56
[ ] #96
[x] #97
[x] Document the testing architecture (see #94)
[x] #59
[x] #61
It might be worthwhile to use an existing testing framework like GoogleTest, but I do not yet know enough about those to make an informed decision.
While working on #23, I have run into some challenges because of our messy testing setup. The property-based testing approach seems to work well but there are a few things that need to be improved:
It might be worthwhile to use an existing testing framework like GoogleTest, but I do not yet know enough about those to make an informed decision.