Rails - Mar 2005 - Semantics of push_with_attributes

In trying to add timestamp support to habtm, I''ve stumped on a question
which I hope you all help me solve.
And that is, when adding a record twice to the list of associations with 
push with attributes, should the following test succeed or not?

def test_pushing_twice
   dev = Devel.new("name"=>"victor")
   p = Project.new("name"=>"whatever")
   dev.projects.push_with_attributes(p, :joined_on => Date.today)
   dev.projects.push_with_attributes(p, :joined_on => Date.today)
   assert_equal(1,dev.projects.size)
end

That is, should the record be added twice, or else updated with new 
info? And for that matter, should the normal push (or <<) method add 
twice if repeated as well?

I think both views are correct, it''s all a matter of semantics, is it
not?
Let me know your thoughts

Victor

Well, since no one commented, I went ahead and opened a ticket for this 
(http://dev.rubyonrails.com/ticket/843) since I understand it makes no 
sense to add the same object twice to an association. Unless there are 
different join attributes, and still am not sure that this is the case 
as well (preferring an update in this case rather than an additional insert)

I''m waiting on this to be resolved -or clarified- and then
I''ll send the
corresponding patch for #596

cheers

Victor

Victor Jalencas wrote:
> In trying to add timestamp support to habtm, I''ve stumped on a
question
> which I hope you all help me solve.
> And that is, when adding a record twice to the list of associations with 
> push with attributes, should the following test succeed or not?
> 
> def test_pushing_twice
>   dev = Devel.new("name"=>"victor")
>   p = Project.new("name"=>"whatever")
>   dev.projects.push_with_attributes(p, :joined_on => Date.today)
>   dev.projects.push_with_attributes(p, :joined_on => Date.today)
>   assert_equal(1,dev.projects.size)
> end
> 
> That is, should the record be added twice, or else updated with new 
> info? And for that matter, should the normal push (or <<) method add 
> twice if repeated as well?
> 
> I think both views are correct, it''s all a matter of semantics, is
it not?
> Let me know your thoughts
> 
> Victor
> _______________________________________________
> Rails mailing list
> Rails-1W37MKcQCpIf0INCOvqR/iCwEArCW2h5@public.gmane.org
> http://lists.rubyonrails.org/mailman/listinfo/rails