In case it isn't clear... I agree with Ivo's statement: "You should not stub the authentication method in a story. The story should test the whole application framework".
emphasis on "story" Zach On Fri, Mar 7, 2008 at 11:13 AM, Zach Dennis <[EMAIL PROTECTED]> wrote: > On Thu, Mar 6, 2008 at 10:38 AM, Pat Maddox <[EMAIL PROTECTED]> wrote: > > On Thu, Mar 6, 2008 at 7:27 AM, Ivo Dancet <[EMAIL PROTECTED]> wrote: > > > You should not stub the authentication method in a story. The story > > > should test the whole application framework, that way your story might > > > look like this: > > > > > > Given a user from company x > > > When logging in > > > And requesting some page > > > Then the application should do all the stuff it should > > > > > > In the given step you set your host, user name and password and you > > > create that user. > > > In the when steps you make the login request and then you can start > > > doing the stuff you really want to test here as you'll have the > > > session you want at that moment. > > > > That's a little too dogmatic for me. You needn't go through the > > entire stack every story, and you needn't go through the entire app > > path either. > > > > IMO you should be simulating as much as possible what the user is > doing. If they have to go through path a, b, c ... to get to Z, then > so be it. This involves keeping a well-factored set of helpers which > can be used to navigate the app just like a user would. > > The only things you probably should be stubbing in a story are third > party components your site may have to integrate with that aren't > realistic to be run for every story (for the most part, there are > other exceptions as well I know). I would avoid stubbing core > functionality related to your site, > > -- > Zach Dennis > http://www.continuousthinking.com > -- Zach Dennis http://www.continuousthinking.com _______________________________________________ rspec-users mailing list rspec-users@rubyforge.org http://rubyforge.org/mailman/listinfo/rspec-users
