Proposal: split out Django support

[noirbizarre]

Hi !

For many reason, it would be better to have Django support splitted out in its own module:

• MongoEngine is a framework agnostic ORM
• Django support is the only framework supported in mongoengine while others have their own module (ex: flask-mongoengine)
• Having to test Django versions make the testing very hard and long (generates 52 combinations without PyMongo 3), it would be easier as a separate project
• separation will allow to update django-mongoengine on new mongo release without the need to touch mongoengine
• Separation of concerns

So, I think it would be better to have:

• django-mongoengine a Django specific support where it would be easier to test on different Django versions
• mongoengine, the framework agnostic ORM where it would be way easier to test on different versions of PyMongo, without testing on Django

Apparently the work was started but not updated : https://github.com/MongoEngine/django-mongoengine

[noirbizarre]

I don't know the state of the current django-mongoengine.
I'm not using Django anymore but I still can help for the transition.

[rozza]

👍 Would make supporting django simpler and we can specifically support different versions.

[rozza]

mongoengine/django-mongoengine is free for use - it had some initial momentum but its been two years since the last commit. It also has a warning not to clone on it. I could rename and add a new repo incase anyone relied on the code.

[mmellison]

👍

I like the idea of re-writing the Django support as its own separate extension. I believe it help the Django integration more coverage and attention.

[mmellison]

There probably needs to be a discussion around if we re-write the Django support, what do we do with the open issues / PRs surrounding Django.

One idea is to label them with a temporary "New Django" label and close them out. Then we can go back and inspect them as needed during development of the new extension.

[noirbizarre]

@seglberg: the label+close seems to be a good solution. It will helps for the bug triage.

For the legacy django-mongoengine, PyPI shows there was no download at all:
https://pypi.python.org/pypi/django-mongoengine

[mmellison]

Okay, I will spend some time labeling any currently open Django bugs.

Is everyone still OK with removing Django support from the MongoEngine "core" repository and re-visiting support with a new, but separate module?

@thedrow @MRigal @DavidBord @rozza @yograterol

I like the whole maintainer voting idea since we have been keeping most of our maintainer communication limited to GitHub :)

+1 from me.

[MacMaru]

+1 we moved to mongoengine because we wanted to be rid of Django.

[rozza]

+1 but we should move the current code to the django mongoengine repo and publish it to pypi - that way no one loses any code and hopefully then some Djangonauts will want to maintain it and make it better :)