First, sorry for the long post, I tried to condence everything here.
I am trying to customize the way children are added to a concept in the editor. E.g, concept A has 0..n children of concept B. In the editor I can make a children list, and users can add new children with the list. In this scenario, I have the following questions:
1) why can't I use menu parts for children in a multiple relationship (0..n, 1..n)? I tried adding custom menu parts but it only worked if the relationship is 0..1 or 1.
2) what is the difference between adding the first element and adding additional elements? When the list is empty there's a placeholder (<<...>>) that I can use to press Crtl-Space and get a list of possible completions (if the target is abstract, I get a list of subtypes; if it has a smart reference, I get a list of possible targets, etc). But after I add the first element I lose this placeholder and all I can do is add new elements and then fix them (unless the target is abstract, in which case I get an <abstract> placeholder which works the same way as <<...>>). So, can I control how the additional elements are chosen? (I know I can control how they are instantiated with the factory, but I am talking about choosing if pressing <enter> after the last item creates a new item or a placeholder like <<...>>). Also, the first element never triggers the construtor of the behavioural model, whereas the additional elements do.
3) how exactly do smart references add to the completion menu? More specifically, concept B (from the example above) has a smart reference (cardinality 1) to C and B also implements IVisible from the base language. In the initial placeholder (<<...>>) for adding Bs to an A, the completion menu lists the possible targets of the smart references, i.e., instances of C. However, it also adds the possible IVisible values (private, protected, public). Shouldn't only the smart references be added to this menu? Is this a bug?
Thanks for the support!!