Adam Johnson

Fair enough. I think you’ll want to fix the www subdomain first, though.

Yes, this seems like a good feature to add upstream first. I would propose it extending `assertNumQueries` to allow `using` to accept both an iterable of database names, like `using={"default",...

A few things need discussing/doing: 1. Don't merge this until django/django#6294 is merged 😄 2. This currently breaks output. Pytest-django diverges from Django test runner by implementing the database as...

Updated now I've updated my Django PR. I still haven't added a config option to disable it or a changelog note, but it's working the same way Django 1.11 will.

No, rebased on latest master and pushed because Travis had lost the logs. Will look at this agian, thanks for nudge

I have (finally) had another look at this in the Djangocon EU 2018 sprints. Here's the status: * It works(*) on plain pytest, but the output comes without colours and...

> In general I think that running this separately in CI is better than with (possibly) every pytest invocation. Checks take milliseconds and can save a lot of painful debugging...

Django’s `string_if_invalid` feature is flawed in many ways, as I covered [in this blog post](https://adamj.eu/tech/2023/08/09/django-perils-string-if-invalid/). Attempting to fix one case here is probably not worth it and could lead to...

I think the fixture will make tests too verbose, and the deprecation will cause unnecessary churn. I think it would be better with a shorter fixture name and a retention...

> ow sure about how the wrapper would work without frame hacking.. can you provide an example? Sorry, no, I can't. I hadn't grasped the whole problem there. > What...