[news.eclipse.modeling.mdt.uml2.ocl] Re: How to validate uniqueness of c

["Kenn Hussey" <khussey@xxxxxxxxxx>]

Note also that there is a "allOwningPackages()" operation defined on 
NamedElement which may be more pertinent in cases where you're concerned 
only with packages.

Kenn

"Christian W. Damus" <cdamus@xxxxxxxxxx> wrote in message 
news:eivr8r$kgp$1@xxxxxxxxxxxxxxxxxxxx
>
> Hi, Andy,
>
> I'm glad it works!  And, yes, allInstances() is very powerful.  It 
> computes
> the set of all instances of a metaclass (including its specializations) in
> the model (in EMF terms, the resource).
>
> allNamespaces() is defined by the spec as an "additional operation".  You
> won't find it in the class diagrams, but under the "Additional Operations"
> heading (below "Constraints") of the NamedElement description, section
> 7.3.2.  The main purpose of these additional operations is to serve in the
> formulation of constraints  :-)
>
> And that's a good point about self.base_Package ... I forgot that you're
> working in a stereotype context.
>
> Cheers,
>
> Christian
>
> Andreas Maier wrote:
>
>>
>> Works like hell ... OCL is so sweet :-) (I used
>> "includes(self.base_Package)" instead of "includes(self)", to be more
>> portable.)
>>
>> I was not consciously aware yet of the context propagation used several
>> times here, nor of the possibility to start an expression with a
>> metaclass name. This opens some new perspectives for me. Thanks for this
>> example!
>>
>> One more question: I did not find allNamespaces() in the OCL 2.0 spec.
>> Is this an extension in the OCL of Eclipse ?
>>
>> Andy
>
> <snip>
>