rspec

Comments and changes to this ticket

• 

  David Chelimsky December 8th, 2007 @ 08:06 PM

  • → Assigned user changed from “” to “Dan North”
  • → State changed from “new” to “open”

  Actually, the whens and thens are stored separately so that they can use the same step names but mean different things in the different contexts.

  Dan - I'm going to throw this back to you again. Do you think we should be able to use the same step with different implementations across step types?

• 

  Dan North December 9th, 2007 @ 11:07 PM

  (Hoping that replying to a lighthouse email adds a new comment)

  We're at slightly cross purposes here. You are correct when you say events

  and outcomes ("whens" and "thens") are stored separately, so that "When I

  eat cheese" and "Then I eat cheese" can be implemented differently. I'm ok

  with this, and in fact I think it's quite useful.

  The thing I want to avoid is having "When I eat cheese" mean different

  things in different scenarios - by allowing multiple instances of "When I

  eat cheese" to have blocks. As I said, this is almost always an accident and

  should cause a failure rather than replacing the previous block associated

  with the step.

  If not, the order you run scenarios becomes significant. Imagine a third

  scenario with "When I eat cheese" with no block. Its author is assuming a

  (specific) existing implementation of the step, and is almost certainly

  unaware there is more than one. This is a Bad Thing.

  So in summary:

  • using the same description for different types of step (say givens and

  outcomes) is fine - and each requires an implementation.

  • multiple implementations of the same type of step with the same

  description is bad.

  Hope that clarifies things. And that lighthouse works how I expect :)

  On Dec 8, 2007 7:06 PM, Lighthouse wrote: