[rspec-users] New rescue_from handling in rspec-rails 1.1.12

Chris Kampmeier chris at kampers.net
Thu Jan 15 15:10:09 EST 2009

On Wed, Jan 14, 2009 at 7:30 PM, David Chelimsky <dchelimsky at gmail.com>wrote:

> On Wed, Jan 14, 2009 at 8:23 PM, Chris Kampmeier <chris at kampers.net>
> wrote:
> > (Or, here's a syntax-highlighted pastie: http://pastie.org/361114)
> Don't forget that the spec should come first :) Also, you're spec'ing
> behaviour, not implementation. So for me, what you've got is this:
> specification: response.should redirect_to("/")
> behaviour: response.should redirect_to("/")

Thanks, definitely helpful. I think the main thing I needed to get in my
head is that testing Rails controllers with rspec-rails is more about
"behavior," unlike a typical unit-style, in that you're really testing the
whole controller stack, and not just an isolated method. It's a little more
high-level, since you really need the whole setup (request, response, etc.)
to get anything done.

> > Anyway, what should I really be doing here? I could use shared examples
> to
> > maintain DRY (it_should_behave_like :access_denied);
> Personally, I've grown weary of it_should_behave_like. I'd write a
> macro so you could say something like:
> describe PostsController do
>  context "handling DELETE /posts/2" do
>    denies_access_when "the post isn't destroyable by the current user" do
>       post = mock_model(Post, :destroyable_by? => false)
>       Post.stub!(:find).and_return(post)
>      post.should_not_receive(:destroy)
>     end
>  end
> end

Ooh, that's interesting. I see how you would implement that, but I'd never
thought of using a "macro" block helper like that, myself. More generally,
could you post a bit on why you've stopped using shared examples, and what
it is about the syntax above that resonates with you? Looking at the two
alternatives, it_should_behave_like seems like a more natural approach to
me--it seems very declarative, and clear--but that might just be because I'm
used to it.

Thanks again!
Chris Kampmeier
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/rspec-users/attachments/20090115/72909d4d/attachment.html>

More information about the rspec-users mailing list