Add deep unordered for nested data structures

[arminfriedl]

Automatically applies unordered to nested elements in complicated data structures.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (9efae14) to head (46c7512).
Report is 3 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff            @@
##            master       #17   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            1         1           
  Lines           93        99    +6     
  Branches        21        25    +4     
=========================================
+ Hits            93        99    +6     

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

[utapyngo]

@arminfriedl, what will you use this for? Can you give a concrete example?

[arminfriedl]

Yes sure - we compare serialized database states (query results) with expected database states (query results). They sometimes contain nested dicts-with-lists-of-dicts-... (you get the idea). There are many of these were there is no "order by" in the query hence order is not really part of the assertion. And the order also does happen to change. We could apply current unordered manually to the expected data, but it is quite cumbersome to do it fine grained in deeply nested structures.

Another case are integration tests for REST responses where we have similar issues. Even though the JSON data contains lists-of-dicts-of-dicts-of-lists-... (again you get the idea), the order of the elements is sometimes irrelevant and an implementation detail which may (and does) change.

Right now we have this helper function to apply unordered recursively in our own code. I do think however the better place for this is upstream. At the end one can choose to use the fine-grained regular unordered or the coarse-grained, recursive unordered_deep, depending on the test case and the data structure.