Thanks Ant, I gave it a try and it works.
About if we fix this or not. As the contribution's artifact already resolved the model, I do not see why we are returnning a dummy model on the contribution. I wonder what is the purpose to have a dummy composite at contribution level. Is it because we want to support a scenario that the contribution artifacts may be resolved lazily (e.g. when calling getArtifacts) , however we may need to know the composite's name? If that is the scenario we want to support, maybe we can add an attribute in the loadContribution method to indicate if we want to resolve the model or not. And in any cases, I would like to see as soon as the artifact is resolved in the contribution, the contribution deployable should return the artifact's model, so we will not have un consistent . I tried to reload the composite from contribution after retrieving the artifacts, that composite is still the dummy one. Yang
