[rspec-users] should_receive test unexpectedly failing

[David Chelimsky]

On Jul 22, 2010, at 10:48 AM, ericindc wrote:

> I'm unsure why this test is still failing.  I've tested it in console
> and the code works as expected.  Is there something wrong with how I
> have my test written?
> 
> ****************** error
> 
> 1) Job selecting the next job to be processed should lock the next job
>    Failure/Error: job.should_receive(:update_attribute).with(:locked,
> true)
>    (#<Job:0x105496488>).update_attribute(:locked, true)
>        expected: 1 time
>        received: 0 times
> 
> 
> ****************** job_spec.rb
> 
>    it "should lock the next job" do
>     job = Factory(:job, :locked => false)
>     job.should_receive(:update_attribute).with(:locked, true)
>     Job.next
>    end
> 
> 
> ****************** job.rb
> 
>    def self.next
>      job = incomplete.unlocked.prioritized.limit(1).first

This job object is not the same object as the on in the spec. To get this to work, you'd have to stub like so:

  Job.stub_chain(:incomplete, :unlocked, :prioritized, :limit).and_return(job)

Personally, I wouldn't bother, especially since you're using a real db backed object anyhow (Factory(:job ...)). I'd just write the spec like this:

job = Factory(:job, :locked => false)
job.next
job.should_not be_locked

Much cleaner and to the point, and less invasive (and therefore less brittle).

HTH,
David

>      job.lock! if job
>      job
>    end
> 
>  def lock!
>    update_attribute(:locked, true)
>  end
> 
>  def unlock!
>    update_attribute(:locked, false)
>  end
> _______________________________________________
> rspec-users mailing list
> rspec-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rspec-users