> Not necessarily. In the case of a stub on a real object, the purpose is to > control the environment in which the example runs. Consider a method on an > object that returns one value before noon and a different value at noon and > after. In an example for another object that depends on the time-dependent > object, we might stub that method to respond as though it were before noon in > one example, and after noon in another. This means that when you run this > spec in the morning, the example that stubs morning-like behaviour is not > changing anything or needing to fail. Same for the other example in the > afternoon. > > Cheers, > David
This is true, but the "morning" spec would fail in the afternoon (or/ and the afternoon spec would fail in the morning). If the stub is not being touched, it should fail *at some point in time*. If the stub is not being touched, yet your specs are passing, then technically there is nothing wrong with your app. _______________________________________________ rspec-users mailing list rspec-users@rubyforge.org http://rubyforge.org/mailman/listinfo/rspec-users
