[#2288] Handle None questions in case detail view

[pi-sigma]

Taiga: #2288

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.06%. Comparing base (783df9c) to head (5b6d746).

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #1132   +/-   ##
========================================
  Coverage    95.06%   95.06%           
========================================
  Files          940      940           
  Lines        33155    33169   +14     
========================================
+ Hits         31518    31532   +14     
  Misses        1637     1637           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Bartvaderkin]

This mock is not accurate and doesn't seem to test the modification.

It makes retrieve_objectcontactmomenten_for_zaak() return a list with a single None value.

While the code change doesn't handle that. It loops over the objectcontactmomenten as if they can't be null and then does a null check on an attribute of an object in that list.

[pi-sigma]

The modification ensures sure that object.contactmoment is not None, so that contactmoment.registratiedatum further down doesn't raise AttributeError. getattr(None, "contactmoment", None) returns None, in exactly the same way that getattr(objectcontactmoment, "contactmoment", None) would return None for an objectcontactmoment that happens to lack a contactmoment. The test will therefore pass with the modification and fail without it. I didn't see a reason to create a fake objectcontactmoment; the mechnism is exactly the same (getattr -> AttributeError -> None).

Also, the code does not assume AFAIK that objectcontactmomenten can't be null; it creates an empty list, calls retrieve_objectcontactmomenten_for_zaak to populate it, then loops over the list.

[Bartvaderkin]

You made a modification that handles if an external sub resource is None, so the test should test with that same subresource being None. And not rely on some random other None value somewhere that happens to make the test pass for some unrelated reason.

With this test we still don't know if you modification works or makes sense, because the situation it handles is not present in the test.