Eclipse Community Forums: EMF

Home » Modeling » EMF » emf load

Wed, 01 April 2009 02:54

Eclipse User

Hi,
I was trying to validate a particular ecore model in emf which gave these
errors.

The opposite of a containment reference must be a containment reference.
A container must have the upper bound of '1' not '-1'

Can you explain me in detail what is the meaning of these errors or under
what condition these errors occur

Re: emf load [message #428872 is a reply to message #428861]

Wed, 01 April 2009 07:22

Eclipse User

gitanjali punj wrote:
> Hi,
> I was trying to validate a particular ecore model in emf which gave
> these errors.
>
> The opposite of a containment reference must be a containment reference.
In other words you must be a child of your mother and your mother must
be a child of you. How did you validate this? Sure you don't mean:

"The opposite of a containment reference must be a *non* containment
reference."?

The negation of this rule is valid, not what you describe.

> A container must have the upper bound of '1' not '-1'
Each EObject is contained in exactly 0 or 1 containers. Check the 'upper
bound' attribute of the reference under question.

> under what condition these errors occur
At least the first one is an error in your validation rules. I doubt
that this was validated with something from www.eclipse.org/emf.

Felix

Re: emf load [message #428873 is a reply to message #428861]

Wed, 01 April 2009 07:23

Eclipse User

Gitanjali,

Comments below.

gitanjali punj wrote:
> Hi,
> I was trying to validate a particular ecore model in emf which gave
> these errors.
>
> The opposite of a containment reference must be a containment reference.
You mean "The opposite of a containment reference must not be a
containment reference" where the above is the opposite of that.
> A container must have the upper bound of '1' not '-1'
>
> Can you explain me in detail what is the meaning of these errors or
> under what condition these errors occur
You can't make both ends of a bidirectional reference be containment.
Containment must induce a tree structure which means one of the sides
must represent the downward direction of the child relationship while
the other must represent the upward direction of the parent
relationship. So the opposite of a containment reference necessarily is
a container reference and a container reference necessarily must have
and upper bound of 1 because you can only have one parent in a tree
structure.

Re: emf load [message #428874 is a reply to message #428872]

Wed, 01 April 2009 07:24

Eclipse User

Felix,

Oops, I'll have to read farther head before replying. :-P

Felix Dorner wrote:
> gitanjali punj wrote:
>> Hi,
>> I was trying to validate a particular ecore model in emf which gave
>> these errors.
>>
>> The opposite of a containment reference must be a containment reference.
> In other words you must be a child of your mother and your mother must
> be a child of you. How did you validate this? Sure you don't mean:
>
> "The opposite of a containment reference must be a *non* containment
> reference."?
>
> The negation of this rule is valid, not what you describe.
>
>> A container must have the upper bound of '1' not '-1'
> Each EObject is contained in exactly 0 or 1 containers. Check the
> 'upper bound' attribute of the reference under question.
>
>> under what condition these errors occur
> At least the first one is an error in your validation rules. I doubt
> that this was validated with something from www.eclipse.org/emf.
>
>
> Felix

Re: emf load [message #428899 is a reply to message #428874]

Wed, 01 April 2009 10:26

Eclipse User

Ed Merks wrote:
> Felix,
>
> Oops, I'll have to read farther head before replying. :-P

Probably you thought 'oh no this guy again, I won't read his post' :-p

Felix

Previous Topic:	Hide/Reorder Nodes/Properties in Editor
Next Topic:	serialized EObjects with xmi

Goto Forum:

-=] Back to Top [=-

Current Time: Sat Jul 19 10:16:32 EDT 2025

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter