Thanks for the PR @eleciawhite

We don't really have a concept of "tests that are known to fail", rather we would either create an issue to track it, or in the case of a difference in behavior to CPython (or unimplemented behavior like this one), we would also create a test in tests/cpydiff that documents the difference. These tests include a small amount of metadata at the top that is used to automatically generate https://docs.micropython.org/en/latest/genrst/index.html

I think this test would be a good candidate for that. Then if we later decide to fix it, the test can be moved into e.g. tests/basics/object_dict.py as part of the fix.