Replies inline, below. On Mon, Jul 18, 2011 at 8:20 PM, Derek Prior <derekpr...@gmail.com> wrote: >> One of the patterns I am trying to >> implement is from Ryan Bates' complex-form-examples[2]. Ryan's example >> doesn't have any tests, so I am trying to write tests for the code. > > He's updated that with a later episode series. Search his site for "nested" > and it's the first hit, I believe.
I've checked this, and I believe his github project to which I referred uses the latest material from the railscast. >> I am starting with a helper function that he wrote, which makes it >> easy to add fields for an association into a form. The method does a >> little reflection, renders a fields_for call into a string, and then >> returns a link_to_function which calls a custom javascript function >> that adds the fields to the form. > > The realization I'm coming to is that a helper spec isn't necessarily the > best thing to cover this type of thing. RSpec is about specifying software > behavior. The behavior you want to test is, if I recall correctly from the > episode, that when the user clicks a link in the page, new association > fields should be added to the form. I would likely skip the helper spec and > test this with a request spec or cucumber scenario using capybara & > selenium. I think this is a very good point. Before I dipped from my cucumber scenario directly into my helper spec, I should have written a request spec, or alternatively made my cucumber scenario more fleshed out. Perhaps what led me down the wrong path was knowing the code I needed to have before I had any tests. Since this is an educational project and there is no deadline to speak of, I have been very strict trying to write the minimal amount of code to make my tests pass, and thus I have very literal tests. In real life, I don't think that is always possible—how would we every write e-mail validators?—but maybe I am over emphasizing it, still. Thoughts? _______________________________________________ rspec-users mailing list rspec-users@rubyforge.org http://rubyforge.org/mailman/listinfo/rspec-users
