| [ATL] Unexpected behavior transforming extended metamodels. [message #481364] |
Thu, 20 August 2009 16:05 |
 Juan Pedro Silva
Messages: 258
Registered: July 2009 |
Senior Member |
|
|
Hello everyone.
I'am having unexpected behavior when trying to transform some extended
metamodel elements (the elements of the extension are not matched as
elements of the superclass in the extended metamodel). I managed to
worked it around with lazy rules, but I'm trying to figure out the
correct way to use extended metamodels in ATL. The story follows.
I created two transformations: WS-PolicyMM <-> MMA, where MMA
practically contains WS-PolicyMM (with very little modifications).
I then have WS-SecurityPolicyMM, which is an extension of WS-PolicyMM in
a similar fashion to "Generating an Extended EMF Model"
( http://help.eclipse.org/ganymede/index.jsp?topic=/org.eclips e.emf.doc/tutorials/slibmod/slibmod.html).
I use a valid (validated through EMF) model created with the editor
generated from those metamodels. I feed this model into the WS-PolicyMM
-> MMA and no problem (instances of WS-SecurityPolicyMM metaclasses are
matched as being sub-types of the WS-PolicyMM metaclasses). I get what
seems to be correct results (I double-check the generated elements with
'oclType()'). The resulting model validates.
However, when feeding this resulting model into the MMA -> WS-PolicyMM
transformation, instances of WS-PolicyMM metaclasses are matched
correctly, but instances of WS-SecurityPolicyMM metaclasses (sub-types
of those in WS-PolicyMM) are not. I'm not placing any restrictions
(condition clauses) in the source pattern.
This happens even when these elements
are part of a precisely-typed collection (the type being a WS-PolicyMM
metaclass), and I guess that the fact that the model validates indicates
those elements correspond to this metaclass.
The elements from the extending metamodel
(WS-SecurityPolicyMM)are not being interpreted as elements of the
super-metaclasses and I don't now why (now they are elements of a whole
different metamodel!!).
Any ideas on which could be the reason?. The behavior is the same both
in ATL 2.0.2 and 3.0.
Regards,
Juan Pedro
|
|
|
Powered by
FUDForum. Page generated in 0.03172 seconds
] 
Search
Help
Register
Login
Home
