@rpkilby thanks for progressing this and getting it done!

Great stuff@rpkilby, and thanks for reviewing@jpadilla!

Hi@Pomax. The best thing to do is to create a failing test case as a PR. You don't have to fix the bug yourself, but a failing test case gives us something to work with.

@rpkilby I've filed#7094, but ran into a bit of an issue in that I can't seem to run the tests for the "between 3.6.3 and 3.6.4" version that this PR lives in. I'll be more than happy to further discuss that in the testcase PR.

I can't seem to run the tests for the "between 3.6.3 and 3.6.4" version that this PR lives in

That's fine. We just need a failing test case against the current master. If you can run the tests locally against v3.6.3 and verify that it succeeds, then that's sufficient.

3.6.3 also seems to not want to get to a point where tests will run... are there additional steps beyond what's in the old CONTRIBUTE.md (e.g. which python version to use for the virtualenv, which django version to peg, etc?) to make the tests run successfully?

(I tried both macos and windows, with python 3.7 and python 2.7 just to see if that was the problem -- it wasn't, django-guardian needs 3.5 or above =D -- with django pegged to 1.11 because this is code from 2017, but nothing seems to get me to a point where the tests will run, unlike running the tests for current master, which worked instantly)

I'm also trying to get this to work on ~3.6.3, but I usually try to run tests via tox. In order to get this to work, I had to unpin the dependencies inrequirements-testing.txt, then I could run:

tox -e py37-django111

Edit:
I'm going to continue this discussion in the PR.

ah, nice. Let me see if I can make that work.

Looks like it's still notquite working (windows gives 58 errors, macos 1) but it's close enough to probably finish up tomorrow morning (Vancouver Island time). Thank you for the help!