A couple of days ago I got really excited about working on the discourse_api that I just jumped right in and started working on a feature that I wanted. I got even more excited as I figured out how things works and was able to get the results that I wanted. That excitement soon ended when I ran the tests for the first time and realized that everything blew up in my face.

I do think that it is good to ride a wave of excitement if you have one, but I think before starting to work on any project you should start by running the tests. The tests I don’t think are entirely broken but to get them to pass you will have to have a very specific environment set up that isn’t really documented anywhere.

So, my goal, before I start writing any more features for the discourse_api is to fix the test suite and document the process so that anybody can quickly run the tests.

One of the things that I need to do is make sure that there is good seed data for the tests to run. I’m not really sure what the best route is for this though. One method is to run a db migration for test data on the discourse database and the other method would be to make sure the disource_api had an admin api key and have it generate all the seed data. Even though the discourse_api depends on discourse I feel like they need to be as decoupled as possible from each other.