Codecov Report

Attention: Patch coverage is 50.63291% with 39 lines in your changes missing coverage. Please review.

Project coverage is 84.29%. Comparing base (2a2fab1) to head (7f456f4).
Report is 2 commits behind head on master.

✅ All tests successful. No failed tests found.

@@            Coverage Diff             @@
##           master    #3550      +/-   ##
==========================================
- Coverage   84.51%   84.29%   -0.22%     
==========================================
  Files         133      133              
  Lines       13855    13885      +30     
  Branches     2930     2929       -1     
==========================================
- Hits        11709    11705       -4     
- Misses       1421     1440      +19     
- Partials      725      740      +15     

@sentrivana yep still working on fixing mypy, will mark as draft until this is ready to be reviewed. I think the cleanest solution will be to replace @ensure_integration_enabled with the old manual checking that the integration is not None whenever we need to access the Integration object, because this allows mypy to recognize that the integration object cannot be None.

We can still use @ensure_integration_enabled anywhere we don't need to access the Integration object

The mypy errors led me to find a couple places where we were not checking that the integration was still enabled, so I think this change will potentially help us avoid making such mistakes in the future

@sentrivana, yeah we are; I found ca. 100 usages after these changes. @ensure_integration_enabled is still helpful anytime we don't need access to the integration object within the patched function, which is quite often apparently. However, it is less useful when we need the integration object, since we anyways need to check whether the integration object is None in order to ensure type safety