Extend DjangoListField to use model queryset if none defined

[jkimbo]

This PR extends the functionality for the DjangoListField to use the default queryset on the Django model if no queryset is provided. This allows it to be used in the top level Query type.

This PR also restricts the types that can be used with DjangoListField to subclasses of DjangoObjectType. I imagine that most people already only pass DjangoObjectTypes to DjangoListField but this just ensures it.

[zbyte64]

I think most would people expect DjangoListField to connect the queryset automatically

[JoaRiski]

Seeing as this PR already has been open for over a month and already has two approvals from eligible approvers, what's the hold-up with merging?

I for one would find this immediately useful when released.

[McPo]

Seeing as this PR already has been open for over a month and already has two approvals from eligible approvers, what's the hold-up with merging?

I for one would find this immediately useful when released.

This is an ongoing issue I have noticed with this project. PRs are getting raised to solve real issues, and are getting approved by 1 or 2 people but not the required 3. As such, It will likely discourage others from contributing.

I would suggest lowering the required amount of approvals, to keep the project moving.

[JoaRiski]

This might be going a bit off-topic of this PR, but @McPo where did you find a mention of 3 approvals being required? I tried looking for PR policies on the CONTRIBUTING.md and other relevant docs without much success. Additionally it doesn't seem like other PRs that have been recently merged have had an issue with just a single review.

[McPo]

This might be going a bit off-topic of this PR, but @McPo where did you find a mention of 3 approvals being required? I tried looking for PR policies on the CONTRIBUTING.md and other relevant docs without much success. Additionally it doesn't seem like other PRs that have been recently merged have had an issue with just a single review.

At the bottom of the PRs it states X amount. Thought it stated more than 2, maybe they changed it? Or maybe Im just plain wrong. Regarding other PRs that I kept stumbling upon, there were several that had reviewers accepting it, but never they never hit the required amount. It wasn't just on this repo specifically, but across graphql-python/ in general. I assumed it might have something to do with the move to 3.0, and they werent overly fussed on merging anymore PR's till that transition was done.

Also to bring it a bit more on topic again (although its already been merged, cheers). I came to this issue due to a similar problem with get_node. Incase this helps anyone else in the future, the issue is #776

[jkimbo]

@McPo yep we changed the number of required reviewers to 2 (I think it was originally 3 some time ago). I agree with you that it should probably be reduced lower and I'll bring it up with the other maintainers when I next get the chance.