Indeed this was my question :)
On 19 mai 07, at 00:20, Tudor Girba wrote:
Now I think I understand the question better. At this time, as we
do not depend on unique names when manipulating data, belongsTo is
no longer a special relationship.
As such, it is no longer mandatory from the point of view of the
execution of the browser, although to preserve the semantics of
analyses that require a namespace for a class, it is better to
create an unknown namespace when you do not have one and add all
the stub classes to that one.
I hope I got it right this time :)
On May 18, 2007, at 9:53 PM, Stéphane Ducasse wrote:
> Sure this is not what I implying.
> I was wondering if we can import a class or entity requiring a
> belongsTo and that this field is empty (most of the time
> this is true with stubs)
>> Hi Stef,
>> The convention is that a root namespace has a nil "belongsTo". In
>> any case, "belongsTo" has no relationship with "stub".
>> On May 18, 2007, at 3:01 PM, Stéphane Ducasse wrote:
>>> I do not remember :)
>>> and in the ST extractor I reified all the namespaces (and forgot
>>> is happening for the root).
>>> Moose-dev mailing list
>> Tudor Girba
>> "To lead is not to demand things, is to make them happen."
"Being happy is a matter of choice."