| Diagram extensions hiding each other? [message #1720404] |
Mon, 18 January 2016 08:09  |
Eclipse User |
|
|
|
Hi,
I'm experimenting with a diagram showing the internals of a container.
These internals might origin from different languages and I created different viewpoints allowing to visualize them. Each viewpoint extends the default layer of the container and adds children to it.
However, it seems that activating one viewpoints results in hiding the subnodes of viewpoint 2.
So I'm having a main language ML containing a Container. Furthermore, I'm having two languages L1 and L2 each containing possible children of the Container.
I created a main viewpoint MVP containing a diagram definition for visualization of the container. Then I created two additional viewpoints VP1 and VP2 each extending the diagram definition of MVP. Each of them adds an additional layer containing the view definitions of the children. However, enabling both of them shadows one of them. So I'm never capable of seeing all children.
Is this fundamentally impossible or am I making a mistake?
Greetings,
Wilbert
|
|
|
|
| Re: Diagram extensions hiding each other? [message #1721997 is a reply to message #1720412] |
Tue, 02 February 2016 03:03  |
| Eclipse User |
|
|
|
Hi Wilbert,
Le 18/01/2016 15:10, Wilbert Alberts a écrit :
> I created an example.
>
> The attached zip file contains the language (in ML, ML.edit and ML.editor) and three viewpoint specification projects.
>
> ML contains a runtime.launch. Executing this launch configuration results in an eclipse instance with (hopefully) four projects. VP, VP1 and VP2 contain the main viewpoint and the two sub viewpoints. The example project, well, contains the example.
>
> The example project contains a representation file showing C, the container. If both layers are activated only one part is visible (p1). If the VP1Parts layer is deactivated, the other part becomes visible (p2).
>
> Any ideas?
This comes from the use of ContainerMappingImports importing the same
mapping. When you activate both layers, this enables both container
mapping imports but Sirius cannot use both: if you have one layer and
one import mapping, when you activate it, Sirius will keep the same view
and bounds and will update the view regarding the activated description.
But when there is two (or more) at the same "import" hierarchy, it takes
only the first to respect the "same view" contract.
A workaround might be to chain the imports.
>
> Greetings,
> Wilbert.
>
Regards
--
Maxime - Obeo
Need training or professional services for Sirius?
http://www.obeodesigner.com/sirius
|
|
|
|
| Re: Diagram extensions hiding each other? [message #1727064 is a reply to message #1721997] |
Fri, 18 March 2016 10:18 |
| Eclipse User |
|
|
|
Hi Maxime,
As it is Friday, I found some time to explore whether your idea actually would work.
So far I found out that it didn't (likely because I made a mistake).
So my situation is following: 3 VSP projects: S, D and C
S contains a diagram specification containing a layer with a containermapping SS with subnodes SSnn.
D contains an extension of the diagram specification of S. It imports the containermapping SS and states that it wants to reuse the previous subnodes as well. Furthermore it adds another kind of subnode DDnn. So far so good,this works.
Doing the same with C leads to the situation that the containermapping from S is extended by two layers. That is not possible as you indicated.
Instead, I try to create a diagramextension of the diagram extension of D. Within that, I would like to import the containermapping extension from D. However, D does not contain a container mapping. It just contains a container mapping import which can not be imported on its turn.
So I can not chain the extensions, I think.
Or did you have something else in mind?
Greetings,
Wilbert.
|
|
|
Powered by
FUDForum. Page generated in 0.07078 seconds
Home