Bringing django-balanced up to date.

[dmpayton]

Howdy,

I'm very excited about (finally!) using Balanced next month for a project, and I've been doing a bit of hacking on django-balanced to familiarize myself with how things work. I have done a bit of cleanup and organization, and resolved a number of outstanding issues, but have yet to implement the v1.1 API (balanced-python 1.0). I'm opening this ticket in hopes of getting a little guidance on moving forward.

https://github.com/dmpayton/django-balanced/tree/dev/rocket-surgery

Here's an approximation of what I've done so far and what I still want to do:

• Fix how balanced is configured (Context Processors overriding other context? #2, Creating Users in shell returns 401 #3, Fix issue #3 where Account creation in shell returns 401 #4, Setting defaults are injected in the __init__.py #5)
• Support custom user models
• Revamp testing
  • Travis-CI integration (Travis-CI Integration #9) https://travis-ci.org/dmpayton/django-balanced
  • Tox integration
  • Test coverage https://coveralls.io/r/dmpayton/django-balanced
  • Test PEP8
• Code organization
  • Forms in forms.py
  • Admin view logic in views.py
• Disable Credit change form in Admin (Credit Admin lets you change info #1)
• Update to Balanced API v1.1 (balanced-python 1.0)
• Python 3 (in preparation for balanced-python supporting py3k)
  • https://caniusepython3.com/check/50d2c2de-ceba-410d-8489-aaacd4ae7670

My next steps are to get things kickin' with the v1.1 API.

Can I assume that this will be a clean break and that maintaining backwards-compatibility with the 1.0 API is not an issue?

This will involve some database changes, most notably the removal of the Account and the addition of Order and Customer. I would also like to change BalancedResource.uri to BalancedResource.href for consistency with balanced-python. Any problem with including an (optional) dependency on South to manage this? (Django 1.7 will ship with a migration framework, but we need to support <1.7)

I want to maximize the chances of all this being merged in, so please let me know if I'm going down the wrong path.

Thanks!

tl;dr Hi, I did some things! Should I keep doing more things?

[steveklabnik]

hey there! Sorry, I didn't see this until now. Yes, we would looooove a PR to bring this up to date.

[steveklabnik]

Can I assume that this will be a clean break and that maintaining backwards-compatibility with the 1.0 API is not an issue?

Yes, that would be totally fine.

[mjallday]

Hey @dmpayton. Turns out we didn't configure notifications for this project so I haven't realised that anyone was interested or still using this. <redface.jpg>

@steveklabnik and I were just discussing what to do here. this sounds awesome.

[mjallday]

Can I assume that this will be a clean break and that maintaining backwards-compatibility with the 1.0 API is not an issue?

We can just semver it and bump the major version. That should not cause any issues.

[mjallday]

I want to maximize the chances of all this being merged in, so please let me know if I'm going down the wrong path.

This sounds phenomenal. @steveklabnik and I can work with you and we'd love the contribution. I had a quick look at your changes so far and they look fantastic, I don't see any issues with merging it in once we all agree that it's functional.

[dmpayton]

@mjallday @steveklabnik Awesome, thank you both so much!

Would it be reasonable to issue a PR on what's already in place and issue a new release? This way anyone using the v1.0 API can get an update that resolves some of the current issues.

My instinct is to call that 0.2.0, but given the removal of BalancedMiddleware and heavy refactor of the admin, this might need to be 1.0.0. The updates for the v1.1 API would again increment the major version number.

Thoughts?

[mjallday]

i'm totally OK with us doing exactly what you mentioned. let's slate the current implementation for 1.0.0 and freeze that then move on to 2.0.0 which will follow the 1.1 revision of the API.

can you create the pull-request and we can go through and see what, if anything, needs to change to get it merged. thanks!

[dmpayton]

@mjallday Done and done. Since there's a plan for moving forward, as well as a place to discuss my changes at a deeper level, I'm going to close out this issue. Thanks again!