rspec users - Aug 2007 - rspec hangs when used with QtRuby

Hello to everyone

I''m having trouble using RSpec to test code which uses QtRuby (both
version 3
and 4).

I noticed that, after upgrading from RSpec 1.0.5 to RSpec 1.0.8, some of my 
specs which used QtRuby caused spec (or rake, when used from a Rakefile) to 
hang indefinitly. After a bit of investigation, I found that the problem 
arose in the following situation:

* requiring Qt4 / Qt3
* creating a stub which stubs a method called :type

What I found particulary strange is that this happens even if I don''t
actually
use Qt in the spec (directly or not). The following is the simplest example I 
could write which reproduces the problem:

  require ''Qt4''

  describe ''something'' do

    it ''should work passing the :type parameter'' do
      s = stub(''x'', :type => ''anything'')
    end

  end

If I remove the first line, spec exits correctly with the message "1
example,
0 failures". With the first line, instead, it hangs, and I have to press 
Ctrl+C to make it go on (of course, in this case it says: "1 example, 1 
failure"). If I use another mock framework (I tried with flexmock), then 
everythings works. If I replace :type with anything else in the call to stub, 
then everything works.

I tried searching google, this mailing list''s archives and QtRuby
forum, but I
found nothing about this. Am I doing something wrong or is it a RSpec or 
QtRuby problem? I''m asking here because I think it''s caused by
RSpec (or by
my wrong use of it), since it happened when I upgraded RSpec. If the issue is 
related more to QtRuby than to RSpec, please let me know and I''ll ask
in the
QtRuby forum.

Thanks in advance

Stefano

On Aug 25, 2007, at 9:50 AM, Stefano Crocco wrote:
> Hello to everyone
>
> I''m having trouble using RSpec to test code which uses QtRuby
(both
> version 3
> and 4).
>
> I noticed that, after upgrading from RSpec 1.0.5 to RSpec 1.0.8,  
> some of my
> specs which used QtRuby caused spec (or rake, when used from a  
> Rakefile) to
> hang indefinitly. After a bit of investigation, I found that the  
> problem
> arose in the following situation:
>
> * requiring Qt4 / Qt3
> * creating a stub which stubs a method called :type
>
> What I found particulary strange is that this happens even if I  
> don''t actually
> use Qt in the spec (directly or not). The following is the simplest  
> example I
> could write which reproduces the problem:
>
>   require ''Qt4''
>
>   describe ''something'' do
>
>     it ''should work passing the :type parameter'' do
>       s = stub(''x'', :type =>
''anything'')
>     end
>
>   end
>
> If I remove the first line, spec exits correctly with the message  
> "1 example,
> 0 failures". With the first line, instead, it hangs, and I have to  
> press
> Ctrl+C to make it go on (of course, in this case it says: "1  
> example, 1
> failure"). If I use another mock framework (I tried with flexmock),  
> then
> everythings works. If I replace :type with anything else in the  
> call to stub,
> then everything works.
I''m not sure which stubbing framework you are using when you
aren''t
using flexmock, but I believe that if you are using RSpec''s default  
mock/stubbing framework, the standard stubbing method is "stub!", not
"stub", although this should still raise an error.  I would suggest  
adding relevant version numbers (of Ruby, Qt, RSpec (gem or trunk),  
platform, OS), and submitting a bug report on Rubyforge''s database if  
no one gets back to you on this one.

Scott