On Jan 10, 2008, at 8:27 PM, James Deville wrote:
> We have a custom implementation of the Mother Object idea. It's inside
> of a module, basically like this:
>
> module Factory
> %w(account friendship person invitation message asset email_address
> birth).each do |klass|
> eval <<-EOF
> def self.create_#{klass}(attributes = {})
> default_attributes = valid_#{klass}_attributes
> #{klass.camelize}.create!
> default_attributes.merge(attributes)
> end
> EOF
> end
>
> def valid_message_attributes(options = {})
> {
> #some message options
> }
> end
>
> #more valid_x_attributes methods
> end
>
> When I require this file in a spec file, Factory.create_person works
> fine. When I try to do it in a step file, Factory.create_person is
> defined, but it fails because valid_person_attributes is missing. If
> I'm in a debugger, and I do require 'model_factory' followed by
> Factory.create_person, same deal.
>
> Do we have a blank slate being loaded somewhere, and why is it wiping
> out require files?
Actually, though, I think John D. Hume is right (and this is a
current bug with FixtureReplacement, as well): your valid_* methods
are not class methods, but instance methods.
I bet the reason that it has worked in the past is because rspec does
funny things with Module#include, and without having the time to look
at the source, I'd bet that rspec ends up delegating to the class
level (but story runner doesn't) - giving you the appearance that
your software works, but in fact is just taking advantage of rspec's
implementation.
Prove me wrong, if you can.
Best,
Scott
_______________________________________________
rspec-users mailing list
rspec-users@rubyforge.org
http://rubyforge.org/mailman/listinfo/rspec-users
