However, the friendly competition has definitely driven RFacebook to finally get its sorely-missing unit tests:

http://www.livelearncode.com/archives/26

Please feel free to check out the new plugin, and see what you think. I hope it puts RFacebook in better standing with the Rails community, and perhaps scratches that itch for other developers.

Interesting, I’d consider unit tests to be the *first* major missing piece. I agree with Chris’ approach here and that RFacebook has significant weaknesses.

The ideal Ruby/Rails/Facebook solution still isn’t out there, so I don’t see a problem with a little friendly “competition” based on two different approaches. It will help the whole Ruby community.

Keep on truckin’ test-first, Chris.

One of the issues for me was the fact that I needed to add Facebook functionality to an existing app, with an existing userbase (many of whom wouldn’t want to use Facebook). For that, a lightweight solution seemed more appropriate than one that allow for desktop usage, console etc. Especially as the code I’ve got is now (wasn’t to begin with though!) fairly well covered with tests

It actually surprised me how straightforward the library was to write, once I sat down to spec it out with tests — but then that’s the beauty of Ruby. In writing the app (now done, currently being tested), I’ve had to add virtually nothing more to it.

Anyway, good luck with RFacebook, and good luck with the tests. I’ll try to keep up to date with it, and will do occasional posts here as and when I think of things worth posting — planning on doing one about how I solved the Observer/ActiveRecord/feed.publishActionOfUser/profile.getFBML issue.

I think the complexity of the RFacebook session classes is perhaps due to the fact that it is actually intended to be able to be used outside of Rails, even in commandline apps (hence the FacebookDesktopSession class). It was not, however, ported directly from PHP – it was definitely written with Ruby in mind, although I was still fairly new to Ruby at the time

I’m planning on checking through your unit tests to see if they can be added on to RFacebook. Thanks for the permission.

On a side note, if you’re developing a Canvas app in the future, you may want to give RFacebook another try (after we get unit testing, of course). There are some great features for Canvas apps that automatically rewrite urls for link_to, redirect_to, image_path, as well as automatically enable usage of session and flash.

Thanks for dropping by. I’m really not trying to do “yet another” Ruby/Rails library for Facebook (in part, that’s why I haven’t wrapped it us as a plugin). I just needed (for my own piece of mind, if nothing else) to have a library with tests.

I originally planned to write tests for RFacebook, and then after looking at the code realised I didn’t understand what was going on as far as the request/response cycle goes. It all seemed very complicated.

The brief library I did (which I’m now using without any probs) came from the opposite direction you took with RFacebook (I’m not saying what I did was better, just that I felt more comfortable with it) — rather than rewriting the php code in Ruby, I wrote the tests, then wrote fresh Ruby to implement the behaviour the tests described.

In the end it was a lot easier than writing the tests for an existing library, but if there’s anything in there (tests or otherwise) that’s useful to RFacebook do go ahead and use it — I’ve released it under the same license as Rails.

Seeing as you’ve gotten some experience writing up unit tests for the Facebook request/response cycle, it would be nice if you could contribute that to the RFacebook project. I’ve requested someone with Rails experience at the new homepage:

http://rfacebook.rubyforge.org

There is a great community of people using the library and vetting the corner-cases, and the lastest version (0.9.1) is quite solid. The last major missing piece is, as you correctly observed, the lack of unit tests. Rather than rolling “yet another” Ruby and Rails library for Facebook, the community would really appreciate if you contributed to the RFacebook library directly. The project is open-source just for that reason Shoot me an email if you’d like to join.