hexgnu
Faraday connection
What
The internals of the linkedin gem are in need of some refactoring. This changes the internals to rely on Faraday and FaradayMiddleware to reduce some of the cruft.
The tests pass but I'm weary as to whether this is good enough yet or not.
One thing that I added was the ability to add middleware to linkedin so that for instance you could do something like this
LinkedIn.middleware << FaradayMiddleware::Caching
@hundredwatt @cmhobbs would love some second eyes on this.
@hexgnu I removed the fail-fast option we set in .rspec (https://github.com/hexgnu/linkedin/commit/aba9e3a8869ef03fb6ba56cc54b4bbfadb1bbbff)
The specs were failing for me while using MultiJSON 1.0.x. I think we'll need to raise that dependency before releasing this.
Were you thinking of releasing this as part of a 0.5 release?
Other than that, I ran my new specs from https://github.com/hexgnu/linkedin/pull/211 against this. They passed when I ran only the new people_spec.rb file, but when I ran the whole suite they failed. I need to track down where the global state failed.
yea I think releasing this as a 0.5.x would be good. And once we get around to introducing OAuth2 we can bump it to 0.6.
Looks like the MultiJSON interface change happend at 1.3.0: https://github.com/intridea/multi_json/commit/e90fd6cb1b0293eb0c73c2f4eb0f7a1764370216, I'll update the gemspec accordingly
