On Wed, Jan 30, 2013 at 10:06 AM, Myron Marston <[email protected]>wrote:

> On Jan 30, 12:33 am, [email protected] wrote:
> > Awesome answer.
> > I have being testing the Given/When/Then framework but the Then prints
> out
> > very ugly so I do not think I will use it.
> >
> > Your solution is good, I thought that doing several expectations inside
> an
> > 'it' was considered bad style.
> > But I think I understand the general idea, what I really want to check is
> > the fact that it has only one page and this high level check can be
> > decomposed on these small checks in this case.
> >
> > Thank you very much.
>
> I believe the "one expectation per example" (or "one assertion per
> test") guideline was born as a reaction to a testing style that would
> pile up many unrelated assertions/expectations in a single test/
> example, which made it hard to understand what behavior was failing
> when you get failure output.


Separate from whether the expectations are related, the guideline is
about understanding context. If 4 examples with one expectation in each all
fail at the same time, you have 4 data points which provide insight into
the problem. If all 4 expectations are in one example, you only get 1 data
point.

The biggest offences of this, however, are scenarios in which the subject's
state is changed multiple times with expectations at each state.

-- 
You received this message because you are subscribed to the Google Groups 
"rspec" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.


Reply via email to