Reorder condition in get_field_value function

[GreatBahram]

Hello there,

Some projects have their own fields that are not connected to any model. With the previous approach, the code would have broken, but with this order, it will consider those too. What do you think?

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 95.29%. Comparing base (5289482) to head (96d3906).
Report is 17 commits behind head on master.

❗ Current head 96d3906 differs from pull request most recent head fbb1bc7

Please upload reports for the commit fbb1bc7 to get more accurate results.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #652      +/-   ##
==========================================
+ Coverage   95.21%   95.29%   +0.07%     
==========================================
  Files          31       31              
  Lines        1025     1042      +17     
==========================================
+ Hits          976      993      +17     
  Misses         49       49              

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

[hramezani]

Thanks @GreatBahram for this PR.
Could you please:

• Add a test to show what problem you are going to fix
• Add an entry to the changelog file.

[GreatBahram]

Thanks @GreatBahram for this PR. Could you please:

* Add a test to show what problem you are going to fix

* Add an entry to the changelog file.

Hi there, I thought it would be beneficial to check the rel_class attribute first and then validate the one_to_one and many_to_one attributes.

This change was disrupting our codebase. We had several non-database fields that functioned correctly before updating the auditlog. The workaround for us was to set the many_to_one and one_to_one attributes to make sure we don't end up in this newly introduced condition. I wonder if it makes sense to do it this way. what do you think?

[hramezani]

GreatBahram

Agree, that make sense to check the rel_class before. we were not aware of your usecase. that's why I asked for a test case. so, next time if somebody change the condition order, the test will fail.

[hramezani]

Better to create a model as you described in the PR description and test with a real model instead of mocking

[GreatBahram]

Hi there, while working on the test case and investigating Django non-database fields reached this conclusion we should have set many_to_many, many_to_one and one_to_one to None, so I'll close this issue, thanks for the support 🌻