be_none and be_any don't seem to work properly

[iain-18]

Hi all!

I'm testing some named scopes. I use be_all(&:some_predicate) to test
if all the elements in the collection are valid, which works, but when
I try to test the inverse, it fails.

You can see the complete code at: http://gist.github.com/213636

But in short:

subject = User.active
subject.should be_all(&:active) # works

inverse = User.all - subject
inverse.none?(&:active).should be_true # works
inverse.should be_none(&:active) # doesn't work

And also "inverse.should_not be_any(&:active)" doesn't work either

And I cannot figure out why. The RDoc of rspec or rspec-rails don't
mention be_none or be_any, so I would think that this would be
straightforward. Anyone any ideas?

Thanks in advance,
Iain - iain.nl
_______________________________________________
rspec-users mailing list
rspec-users@...
http://rubyforge.org/mailman/listinfo/rspec-users

[David Chelimsky-2]

On Oct 19, 2009, at 2:12 PM, iain wrote:

> Hi all!
>
> I'm testing some named scopes. I use be_all(&:some_predicate) to test
> if all the elements in the collection are valid, which works, but when
> I try to test the inverse, it fails.
>
> You can see the complete code at: http://gist.github.com/213636
>
> But in short:
>
> subject = User.active
> subject.should be_all(&:active) # works
>
> inverse = User.all - subject
> inverse.none?(&:active).should be_true # works
> inverse.should be_none(&:active) # doesn't work
>
> And also "inverse.should_not be_any(&:active)" doesn't work either
>
> And I cannot figure out why. The RDoc of rspec or rspec-rails don't
> mention be_none or be_any, so I would think that this would be
> straightforward. Anyone any ideas?

I'm not certain, but I _think_ that the problem is the way you're  
setting up the factories. Try just creating the objects directly and  
see what happens:

http://gist.github.com/215201

HTH,
David



>
> Thanks in advance,
> Iain - iain.nl
> _______________________________________________
> rspec-users mailing list
> rspec-users@...
> http://rubyforge.org/mailman/listinfo/rspec-users

Cheers,
David



_______________________________________________
rspec-users mailing list
rspec-users@...
http://rubyforge.org/mailman/listinfo/rspec-users

[iain-18]

On Oct 21, 5:45 pm, David Chelimsky <dchelim...@...> wrote:

> On Oct 19, 2009, at 2:12 PM, iain wrote:
>
>
>
> > Hi all!
>
> > I'm testing some named scopes. I use be_all(&:some_predicate) to test
> > if all the elements in the collection are valid, which works, but when
> > I try to test the inverse, it fails.
>
> > You can see the complete code at:http://gist.github.com/213636
>
> > But in short:
>
> > subject = User.active
> > subject.should be_all(&:active) # works
>
> > inverse = User.all - subject
> > inverse.none?(&:active).should be_true # works
> > inverse.should be_none(&:active) # doesn't work
>
> > And also "inverse.should_not be_any(&:active)" doesn't work either
>
> > And I cannot figure out why. The RDoc of rspec or rspec-rails don't
> > mention be_none or be_any, so I would think that this would be
> > straightforward. Anyone any ideas?
>
> I'm not certain, but I _think_ that the problem is the way you're  
> setting up the factories. Try just creating the objects directly and  
> see what happens:
>
> http://gist.github.com/215201
>
> HTH,
> David
>
>
>
> > Thanks in advance,
> > Iain - iain.nl
> > _______________________________________________
> > rspec-users mailing list
> > rspec-us...@...
> >http://rubyforge.org/mailman/listinfo/rspec-users
>
> Cheers,
> David
>
> _______________________________________________
> rspec-users mailing list
> rspec-us...@...://rubyforge.org/mailman/listinfo/rspec-users

Nope, no difference.
The arrays with records are all fine when I inspect them, just as I
would expect.
The message "expected any? to return false" looks okay, so the only
thing I can imagine is that it doesn't send it to the right receiver.

I've made an inverse selection too, where I match be_any on subject,
but to no avail.

I've also tried it without named_scopes, again not changing the
situation.

But then again, why should "all?" work, but "any?" or "none?" not?
_______________________________________________
rspec-users mailing list
rspec-users@...
http://rubyforge.org/mailman/listinfo/rspec-users

[Tom Stuart]

On 21 Oct 2009, at 16:45, David Chelimsky wrote:
> On Oct 19, 2009, at 2:12 PM, iain wrote:
>> I use be_all(&:some_predicate) to test if all the elements in the  
>> collection are valid, which works, but when I try to test the  
>> inverse, it fails.
> I'm not certain, but I _think_ that the problem is the way you're  
> setting up the factories.

I don't think be_all(&:some_predicate) actually works at all, except  
in the accidental sense that Enumerable#all indiscriminately returns  
true as long as all of the entries in the collection are truthy.

At the moment it looks like dynamic predicate matchers like be_all  
completely ignore their block argument (&:some_predicate doesn't get  
rolled into *args) so you just get the default blockless behaviour of  
the underlying predicate.

I've reported and patched this on Lighthouse: https://rspec.lighthouseapp.com/projects/5645-rspec/tickets/905

Cheers,
-Tom
_______________________________________________
rspec-users mailing list
rspec-users@...
http://rubyforge.org/mailman/listinfo/rspec-users