Eclipse Community Forums RDF feed

Re: [Eugenia] Subdiagrams

Christoph Wienands — 2009-11-20T03:13:14-00:00

As I just got a question from someone regarding subdiagrams:

I recently submitted an enhancement request for subdiagrams of same type to Eugenia at https://bugs.eclipse.org/bugs/show_bug.cgi?id=295271. I attached a ZIP file that contains a simple state machine language that features composite states (states are contained in compartments) and partitioning states, which open subdiagrams through double-click on the border (if you click inside the rectangle, you edit the label). This example was derived from the Eugenia state machine example.

The key were two things: First of all, the @diagram class needs to get a @node annotation as well. That enables this class to have a diagram pane representation as well as a node representation inside diagrams. Then you'll need to make one manual configuration to gmfmap. The exact element, property and value is described in the enhancement request. I also provided three lines of code for an Ecore2GMF.eol file to make such a change automatically. Basically, all you have to do is figure out which class has to have both the diagram representation and the node representation inside a diagram.

Re: ECore2GMF.eol: How to access/change node in GmfMap

Christoph Wienands — 2009-11-20T00:50:31-00:00

Never mind. I finally discovered the println feature in EOL and figured out that the nodeMapping.relatedDiagrams attribute is in fact an EolCollectionType object. Now don't ask me how a collection is stuffed into an XML attribute at the file level (seems counter-intuitive to me), but I finally got this code here working:

var nodeMapping := GmfMap!NodeMapping.all.selectOne(nm|nm.domainMetaElement.name = "PartitioningState");
var canvasMapping := GmfMap!CanvasMapping.all.first();
nodeMapping.relatedDiagrams.add(canvasMapping);

Christoph

Re: ECore2GMF.eol: How to access/change node in GmfMap

Christoph Wienands — 2009-11-19T23:53:18-00:00

Hey Dimitrios,

that already got me further. So here is what I have right now:

var nodeMapping := GmfMap!NodeMapping.all.selectOne(nm|nm.domainMetaElement.name = "PartitioningState");
var canvasMapping := GmfMap!CanvasMapping.all.first();
nodeMapping.relatedDiagrams := canvasMapping;

Unfortunately, the last line crashes because the canvasMapping cannot be assigned to the nodeMapping's relatedDiagram attribute. In the gmfmap XML, this is what the output should be:

I confirmed that the referenced @diagram node is truly a CanvasMapping type. I also confirmed that nodeMapping.relatedDiagrams is valid (it works if I assign 'null').

Any idea about how to get the CanvasMapping into the relatedDiagrams attribute?

Thanks, Christoph

PS: If I get this working, I'll add it to the diagram partitioning enhancement request that I submitted.

Re: Model Elemenet Types: "!" vs. "::"

Horacio — 2009-11-19T14:37:33-00:00

Dimitris,

Thanks for your help! I have a final question. On your mail you told me:

Quote:

2. I changed both

to

Why and when to set that property to true, and where can I get detailed information on the attributes of the Epsilon workflow tasks?

Thanks,

Horacio

Re: EVL validation is working, EMF is not

giovanni — 2009-11-19T12:24:44-00:00

Hi Dimitris,

I filed the bug, the link is:

https://bugs.eclipse.org/bugs/show_bug.cgi?id=295575

Regards,

Giovanni.

Re: EVL validation is working, EMF is not

Dimitrios Kolovos — 2009-11-19T10:52:43-00:00

Hi Giovanni,

I've reproduced this and it seems to be a bug. Could you please open a
bug report and post the link here?

Cheers,
Dimitris

giovanni wrote:
> Sure, this is my environment:
>
> - Eclipse Modeling Galileo SR1 Incubation
> - Epsilon (by stable update site)
> - EMFatic
>
> I developed a new GMF Project from an *.emf file:
> @namespace(uri="hierarchy", prefix="hierarchy")
> @gmf(foo="bar")
> package hierarchy;
>
> @gmf.link(source="From",target="To",style="solid",width=
> "1",label.icon="false",tool.small.path="icons\obj16\Association.gif
> ",tool.small.bundle="org.eclipse.uml2.diagram.common")
> class Arc {
> ref Node[1]#arcs From;
> ref Node[1] To;
> id attr String name;
> attr boolean[1] multiple = false;
> attr boolean[1] optional = false;
> }
>
> @gmf.node(figure="ellipse",size="15,15",color="0,0,0",tool.small.path=
> "icons\obj16\ActivityFinalNode.gif",tool.small.bundle="org.eclipse.uml2.diagram.common
> ")
> class Dimension extends Node {
> }
>
> @gmf.node(figure="ellipse",size="10,10",tool.small.path=
> "icons\obj16\Artifact.gif",tool.small.bundle="org.eclipse.uml2.diagram.common
> ")
> class Attribute extends Node {
> attr boolean[1] crossDimensional = false;
> attr boolean[1] descriptive = false;
> attr boolean[1] optional = false;
> attr boolean[1] convergence = false;
> attr boolean[1] shared = false;
> attr boolean[1] recursive = false;
> }
>
> @gmf.diagram(onefile="true",model.extension="hrc",diagram.extension= "hrc")
> class Hierarchy {
> id attr String[1] name;
> val Dimension[1] dimension;
> val Attribute[*] attributes;
> attr String description;
> attr String businessDescription = "";
> attr Date dateLastModified;
> attr Date date;
> attr Type type = "NOT_CONFORMED";
> val FactSchema factSchema;
> }
>
> @gmf.node(label="name",label.placement="external",label.icon= "false")
> abstract interface Node {
> id attr String[1] name;
> attr String businessDescription;
> attr Date date;
> attr Date dateLastModified;
> attr String description;
> attr DataType dataType;
> val Arc[*]#From arcs;
> }
>
> datatype Date : java.util.Date;
>
> enum Type {
> CONFORMED = 0;
> NOT_CONFORMED = 1;
> DEGENERATED = 2;
> }
>
> enum DataType {
> CHAR = 0;
> VARCHAR = 1;
> TEXT = 2;
> INTEGER = 3;
> FLOAT = 4;
> }
>
> @gmf.node(label="name",label.placement="internal",figure=
> "rounded",size="60,90",tool.small.path="icons\obj16\ActivityPartition.gif
> ",tool.small.bundle="org.eclipse.uml2.diagram.common")
> class FactSchema extends Node {
> @gmf.compartment(layout="list")
> val Measure[+] measures;
> }
>
> @gmf.node(label="name",label.icon="false",label.placement=
> "internal",figure="rectangle",size="60,25",tool.small.path=
> "icons\obj16\ActivityParameterNode.gif",tool.small.bundle="org.eclipse.uml2.diagram.common
> ")
> class Measure {
> id attr String[1] name;
> attr String businessDescription;
> attr Date date;
> attr Date dateLastModified;
> attr String description;
> attr boolean ~derived;
> attr String formula = "";
> }
>
>
> and thus I added a validation plugin with an *.evl file with several
> rules, that are respected when I invoke "Validate" from the menu of my
> application. But the implicit constraint derived from EMF model (i.e.:
> at least one Dimensione per Hierarchy, or every Object must have a name,
> ....) are not validated.
>
> Instead, the NullPointerException was resolved, the cause was an error
> in the *.evl file synthax.

--
Spread the word: http://www.eclipse.org/gmt/epsilon/spreadtheword
Follow Epsilon on Twitter: http://twitter.com/epsilonews

Re: EVL validation is working, EMF is not

giovanni — 2009-11-19T08:42:19-00:00

Sure, this is my environment:

- Eclipse Modeling Galileo SR1 Incubation
- Epsilon (by stable update site)
- EMFatic

I developed a new GMF Project from an *.emf file:

@namespace(uri="hierarchy", prefix="hierarchy")
@gmf(foo="bar")
package hierarchy;

@gmf.link(source="From",target="To",style="solid",width= "1",label.icon="false",tool.small.path="icons\obj16\Association.gif ",tool.small.bundle="org.eclipse.uml2.diagram.common")
class Arc {
ref Node[1]#arcs From;
ref Node[1] To;
id attr String name;
attr boolean[1] multiple = false;
attr boolean[1] optional = false;
}

@gmf.node(figure="ellipse",size="15,15",color="0,0,0",tool.small.path= "icons\obj16\ActivityFinalNode.gif",tool.small.bundle="org.eclipse.uml2.diagram.common ")
class Dimension extends Node {
}

@gmf.node(figure="ellipse",size="10,10",tool.small.path= "icons\obj16\Artifact.gif",tool.small.bundle="org.eclipse.uml2.diagram.common ")
class Attribute extends Node {
attr boolean[1] crossDimensional = false;
attr boolean[1] descriptive = false;
attr boolean[1] optional = false;
attr boolean[1] convergence = false;
attr boolean[1] shared = false;
attr boolean[1] recursive = false;
}

@gmf.diagram(onefile="true",model.extension="hrc",diagram.extension= "hrc")
class Hierarchy {
id attr String[1] name;
val Dimension[1] dimension;
val Attribute[*] attributes;
attr String description;
attr String businessDescription = "";
attr Date dateLastModified;
attr Date date;
attr Type type = "NOT_CONFORMED";
val FactSchema factSchema;
}

@gmf.node(label="name",label.placement="external",label.icon= "false")
abstract interface Node {
id attr String[1] name;
attr String businessDescription;
attr Date date;
attr Date dateLastModified;
attr String description;
attr DataType dataType;
val Arc[*]#From arcs;
}

datatype Date : java.util.Date;

enum Type {
CONFORMED = 0;
NOT_CONFORMED = 1;
DEGENERATED = 2;
}

enum DataType {
CHAR = 0;
VARCHAR = 1;
TEXT = 2;
INTEGER = 3;
FLOAT = 4;
}

@gmf.node(label="name",label.placement="internal",figure= "rounded",size="60,90",tool.small.path="icons\obj16\ActivityPartition.gif ",tool.small.bundle="org.eclipse.uml2.diagram.common")
class FactSchema extends Node {
@gmf.compartment(layout="list")
val Measure[+] measures;
}

@gmf.node(label="name",label.icon="false",label.placement= "internal",figure="rectangle",size="60,25",tool.small.path= "icons\obj16\ActivityParameterNode.gif",tool.small.bundle="org.eclipse.uml2.diagram.common ")
class Measure {
id attr String[1] name;
attr String businessDescription;
attr Date date;
attr Date dateLastModified;
attr String description;
attr boolean ~derived;
attr String formula = "";
}

and thus I added a validation plugin with an *.evl file with several rules, that are respected when I invoke "Validate" from the menu of my application. But the implicit constraint derived from EMF model (i.e.: at least one Dimensione per Hierarchy, or every Object must have a name, ....) are not validated.

Instead, the NullPointerException was resolved, the cause was an error in the *.evl file synthax.

Re: [EuGENia]

2009-11-18T12:54:58-00:00

Originally posted by: arendt.mathematik.uni-marburg.de

Hi Dimitris,

Thanks a lot, it works!
I specified in the model:

@gmf.link(source="source", target="target")
class Association {
val Property[2] ownedEnds;
derived ref Classifier source;
derived ref Classifier target;
}

.... and implemented in AssociationImpl.java:

public Classifier getSource(){
return getOwnedEnd().get(0).getType();
}

public Classifier getTarget(){
return getOwnedEnd().get(1).getType();
}

Best regards,
Thorsten

Dimitris Kolovos schrieb:
> Hi Thorsten,
>
> It seems that setting both the source and target to "ownedEnds" doesn't
> work. I'd recommend trying to create two derived features and setting
> them as the source and target of the link respectively:
>
> @gmf.link(source="source", target="target", label="name")
> class Association {
> ref Property[2] ownedEnds;
> derived ref Property source;
> derived ref Property target;
> }
>
> You'll then need to implement the logic in which the features are
> derived from ownedEnds by customizing the generated EMF code - don't
> forget to set @generated NOT in the methods you change. I must admit I
> haven't tried it but it should work. Please let me know if it doesn't
> and I'll have a closer look.
>
> Cheers,
> Dimitris
>
> Thorsten Arendt wrote:
>> Hi,
>>
>> Using EuGENia two questions arose when specifying source and target of a
>> link:
>> 1. How can I specify elements in an EList?
>> 2. How can I navigate to other related EClasses?
>>
>> Consider the following example:
>>
>> class Classifier{}
>>
>> @gmf.node
>> class Class extends Classifier{}
>>
>> class Property extends Classifier{
>> ref Classifier type;
>> }
>>
>> If I want to specify Associations like in UML, I think of this solution:
>>
>> @gmf.link(source=X, target=Y)
>> class Association {
>> val Property[2] ownedEnds;
>> }
>>
>> where X had to be set to "ownedEnds.get(0).getType()" or something like
>> this. Who can help to specify X and Y???
>>
>> Regards,
>> Thorsten
>
>

Re: [EuGENia]

Dimitrios Kolovos — 2009-11-18T12:14:01-00:00

Hi Thorsten,

It seems that setting both the source and target to "ownedEnds" doesn't
work. I'd recommend trying to create two derived features and setting
them as the source and target of the link respectively:

@gmf.link(source="source", target="target", label="name")
class Association {
ref Property[2] ownedEnds;
derived ref Property source;
derived ref Property target;
}

You'll then need to implement the logic in which the features are
derived from ownedEnds by customizing the generated EMF code - don't
forget to set @generated NOT in the methods you change. I must admit I
haven't tried it but it should work. Please let me know if it doesn't
and I'll have a closer look.

Cheers,
Dimitris

Thorsten Arendt wrote:
> Hi,
>
> Using EuGENia two questions arose when specifying source and target of a
> link:
> 1. How can I specify elements in an EList?
> 2. How can I navigate to other related EClasses?
>
> Consider the following example:
>
> class Classifier{}
>
> @gmf.node
> class Class extends Classifier{}
>
> class Property extends Classifier{
> ref Classifier type;
> }
>
> If I want to specify Associations like in UML, I think of this solution:
>
> @gmf.link(source=X, target=Y)
> class Association {
> val Property[2] ownedEnds;
> }
>
> where X had to be set to "ownedEnds.get(0).getType()" or something like
> this. Who can help to specify X and Y???
>
> Regards,
> Thorsten

--
Spread the word: http://www.eclipse.org/gmt/epsilon/spreadtheword
Follow Epsilon on Twitter: http://twitter.com/epsilonews

[EuGENia]

2009-11-18T11:55:08-00:00

Originally posted by: arendt.mathematik.uni-marburg.de

Hi,

Using EuGENia two questions arose when specifying source and target of a
link:
1. How can I specify elements in an EList?
2. How can I navigate to other related EClasses?

Consider the following example:

class Classifier{}

@gmf.node
class Class extends Classifier{}

class Property extends Classifier{
ref Classifier type;
}

If I want to specify Associations like in UML, I think of this solution:

@gmf.link(source=X, target=Y)
class Association {
val Property[2] ownedEnds;
}

where X had to be set to "ownedEnds.get(0).getType()" or something like
this. Who can help to specify X and Y???

Regards,
Thorsten