[news.eclipse.modeling.gmt] Re: [MOFScript]How to write this loop in MOF

[Tor Neple <tor.neple@xxxxxxxxx>]

Hmm again, without the Ecore file it is hard..

In your C class, have you got a reference to the incoming and outgoing 
Objects? If so a forEach on S's outpros should work I guess. From the 
found C instance in each case you should be able to get to the G 
instance by following the links forward.

The key trick is to find out when you are in a S instance, and that 
should be quite simple, you could for instance use if(o.oclIsTypeOf(S)) 
and then start the forEach on the found S instance.

Not sure this helped to much, but I guess it should be possible, again 
all depending on the details of the metamodel

Tor




yeliu wrote:
> Hello,
> When it comes to the metamodel,it's a long story to tell.As I posted
> before,there are four elements(N,K,S,G) which inherit an abstract class
> called Objects.N also contains other Objects including itself.Besides
> there is another class C which connects other four elements and is also
> contained in N. N and K can be connected by only one incoming C and one
> outgoing C. S has only one incoming C and more than one outgoing C.
> Conversely,G can hava more than one incoming C and only one outgoing
> C.So now imagine those elements are connected in the following simple
> sequence:
>
> K
> |
> S
> / | \
> c1 c2 c3
> \ | /
> G
> |
> N
> As shown in the graph,S has 3 outports ,G has 3 inports.
> Now I wanna print them into code,like:
> connect K->S
> connect S.outport[1]->c1->G.inport[1]
> connect S.outport[2]->c2->G.inport[2]
> connect S.outport[3]->c3->G.inport[3]
> connect G->N
> It's easier When iterate elements with only one incoming C and outgoing
> C.I write code like this:
> in the context of N.(objects stands for reference from N to Objects)
> ----------------------------------------
> self.objects->forEach(o:Objects){
> if(o.outgoingC!=null){
> if(!o.oclIsTypeOf(S)){
> o.name ' -> ' o.outgoingC.first().name ' -> '
> }
> ----------------------------------
> but when it iterate to S ,the' connect ' statement should split into 3
> lines, and it's getting hard to get access to objects of C because C is
> not Objects and has different containment relations to N. and when print
> branchs of S,the printed statement should stop at G.
>
> So , generally,the metamodel is like this and my purpose is also
> restated.I don't know if this makes you more confused or a little bit
> clearer.If you still need more detail,just tell me ,thank you for your
> kind help!
>
>
> Ye