Reconfigured the client

[clintonb]

• Inheriting directly from Slumber to avoid having to update the client anytime an endpoint is added/removed.
• Using custom exceptions specific to this client.

@jimabramson This is in response to your comment about exceptions. FYI @rlucioni

[coveralls]

Changes Unknown when pulling eb65fd3 on redux into * on master*.

[clintonb]

Note that the client now becomes quite thin/dumb.

[clintonb]

@benpatterson can you look into the multiple Travis builds issue for this repo?

[clintonb]

@benpatterson also, please disable coveralls comments.

[coveralls]

Changes Unknown when pulling 616969b on redux into * on master*.

[coveralls]

Changes Unknown when pulling 39efb48 on redux into * on master*.

[rlucioni]

@clintonb you can address the multiple Travis builds issue by adding the following to this repo's .travis.yml:

branches:
    only:
        - master

For an example, refer to edx/ecommerce's Travis configuration.

[jimabramson]

@clintonb nicely done. However, this is a more drastic change that I'd envisioned when commenting. I do see the beauty of no longer needing to maintain the client to get access to server methods. But - don't kill me - if we switch over to having developers explicitly interact with slumber's REST api, then rewriting slumber exceptions seems more confusing than helpful. Of course, if the exception-rewriting part were removed there'd be almost nothing left in this client except a wrapper around API.init, which makes me curious what the main intent for separating out the client was.

[coveralls]

Changes Unknown when pulling 4ea334d on redux into * on master*.

[clintonb]

@jimabramson The biggest benefit of the wrapper is removing the need for consumers to understand JWT. The client definitely needs to be separated from LMS so that it can be consumed by both LMS and our acceptance tests. I can go either way on the exceptions.

[jimabramson]

@clintonb thanks, makes sense - in that case I think I would be in favor
of just losing the exceptions stuff.

On Friday, May 1, 2015, Clinton Blackburn notifications@github.com wrote:

@jimabramson https://github.com/jimabramson The biggest benefit of the
wrapper is removing the need for consumers to understand JWT. The client
definitely needs to be separated from LMS so that it can be consumed by
both LMS and our acceptance tests. I can go either way on the exceptions.

—
Reply to this email directly or view it on GitHub
#2 (comment)
.

[coveralls]

Changes Unknown when pulling 54cd879 on redux into * on master*.

[clintonb]

@jimabramson Any opposition to the imports here? Consumers are slightly cleaner since they don't have to import slumber directly.

[jimabramson]

that seems good, yeah.

[jimabramson]

👍

[coveralls]

Changes Unknown when pulling 3e0fe44 on redux into * on master*.

[rlucioni]

👍