Home » Modeling » GMF (Graphical Modeling Framework) » Problem with GMF editor resulting from from xsd
| Problem with GMF editor resulting from from xsd [message #98958] |
Tue, 30 January 2007 10:07  |
Eclipse User |
|
|
|
Originally posted by: christian.daenekas.informatik.uni-oldenburg.de
Hello,
I thought of using a xsd as the base for an editor. So I generated an
ecore and genmodel file from the xsd that we are using in our project.
Everything worked fine for me until code generation. In ecore models
derived from xsd models there is this artificial root element
Diagramroot. The generated code contains the following passage:
/**
* @generated
*/
private static EObject createInitialRoot(DocumentRoot model) {
DocumentRoot docRoot = EditorFactory.eINSTANCE.createDocumentRoot();
docRoot.setnull(model);
return docRoot;
}
The setnull method that is called does not exist and I have no idea what
is supposed to be done there. When I delete the line (thought it´s worth
a try) everything compiles fine, the editor application is able to
create a diagram but I am not able to place any nodes from the palette
onto the Canvas. The file containing the "semantic" model contains
nothing but this:
<?xml version="1.0" encoding="UTF-8"?>
/>
Does anyone here have an idea whats going wrong there? The xsd is valid
so this can´t be the problem I think. Thank you in advance.
Greetings,
Christian
|
|
| | |
| Re: Problem with GMF editor resulting from xsd [message #99565 is a reply to message #98987] |
Wed, 31 January 2007 09:30  |
| Eclipse User |
|
|
|
Please take a look at the screenshot attached. Map is domainMetaElement of
the MindMap domain model, and DocumentRoot.map feature (highlighted) is the
feature I was referring to below. You may want to check
org.eclipse.gmf.examples.mindmap (/model/mindmap.ecore) from GMF examples,
which is xsd-based model.
There's some nasty logic detecting exact feature, perhaps, there's something
wrong with what's expected. Would be great if you provide us with copy of
..xsd file you are using (actually, definition of top xml elements only is
needed, you can wipe the rest).
To workaround the issue, just replace setnull with appropriate setter for
your root model element.
Artem
"Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de> wrote
in message news:epno3r$isa$1@utils.eclipse.org...
> Okay I just found something related to this issue, but I don?t fully
> understand. Where do I have to look for a failure?
>
> From: "Artem Tikhomirov":
>
> 'null' means generator failed to find a feature to set domainMetaElement
> into DocumentRoot. Could you please check there's a stuctural feature
> inside
> DocumentRoot with domainMetaElement as eReferenceType?
>
> Artem
>
> "Jiefei" <majiefei@xxxxxxxxxxxxxxxxx> wrote in message
> news:f89f488bf13cbbcf8d32a03d5e85f3a3$1@xxxxxxxxxxxxxxxxxx
> > Hi,
> > I mapped the DocumentRoot as the doaminMetaElement of the diagram.
> But in
> > the generated diagram code, the createInitialRoot(DocumentRoot model)
> > method (in the FooDiagramEditorUtil class) has the implementation like:
> >
> > DocumentRoot docRoot = FooFactory.eINSTANCE.createDocumentRoot();
> > docRoot.setnull(model);
> > return docRoot;
> >
> > I dont understand why it generates "setnull". is it a bug?
> >
> > cheers,
> >
> > Jiefei
> >
>
>
> Christian Danekas schrieb:
>> Forgot to say that I?m using GMF 2.0 (19.01.2007 nightly).
>>
>> Christian Danekas schrieb:
>>> Hello,
>>>
>>> I thought of using a xsd as the base for an editor. So I generated an
>>> ecore and genmodel file from the xsd that we are using in our project.
>>> Everything worked fine for me until code generation. In ecore models
>>> derived from xsd models there is this artificial root element
>>> Diagramroot. The generated code contains the following passage:
>>>
>>> /**
>>> * @generated
>>> */
>>> private static EObject createInitialRoot(DocumentRoot model) {
>>> DocumentRoot docRoot =
>>> EditorFactory.eINSTANCE.createDocumentRoot();
>>> docRoot.setnull(model);
>>> return docRoot;
>>> }
>>>
>>> The setnull method that is called does not exist and I have no idea
>>> what is supposed to be done there. When I delete the line (thought
>>> it?s worth a try) everything compiles fine, the editor application is
>>> able to create a diagram but I am not able to place any nodes from the
>>> palette onto the Canvas. The file containing the "semantic" model
>>> contains nothing but this:
>>>
>>> <?xml version="1.0" encoding="UTF-8"?>
>>> />
>>>
>>> Does anyone here have an idea whats going wrong there? The xsd is
>>> valid so this can?t be the problem I think. Thank you in advance.
>>>
>>> Greetings,
>>> Christian
|
|
|
| Re: Problem with GMF editor resulting from xsd [message #100617 is a reply to message #99565] |
Fri, 02 February 2007 10:07 |
| Eclipse User |
|
|
|
Originally posted by: christian.daenekas.informatik.uni-oldenburg.de
Thank you, I managed to solve the problem by modifying the ecore model.
Mine looks exactly the same as the Mindmap model concerning the
DocumentRoot Class. Instead of Map I have a reference to differently
named class, which is also the diagram canvas. But there must be
something wrong still. When I generate all the code and launch the
application I can create a new diagram. But sometimes there is no canvas
element created at all and I can´t place my nodes on the Diagram, and
other times there is first the DocumentRoot and the canvas I defined is
below (in the semantic model file). I want my editor start the semantic
model files like in the mindmap example where the DocumentRoot element
can not be seen and the tree structure starts with <mindmap . I can´t
find a difference between my models and the mindmap example. But when I
changed the Extended metadata of my canvas element (the name started
with lower case, I altered it to upper case) the behavior changed and
the canvas plus DocumentRoot was created (before this change there was
no canvas). I fail to see the logic behind this. Could you help me out
again?
Greetings,
Christian
Artem Tikhomirov schrieb:
> <<untitled2.png>>
> Please take a look at the screenshot attached. Map is
domainMetaElement of
> the MindMap domain model, and DocumentRoot.map feature (highlighted)
is the
> feature I was referring to below. You may want to check
> org.eclipse.gmf.examples.mindmap (/model/mindmap.ecore) from GMF
examples,
> which is xsd-based model.
>
> There's some nasty logic detecting exact feature, perhaps, there's
something
> wrong with what's expected. Would be great if you provide us with
copy of
> .xsd file you are using (actually, definition of top xml elements
only is
> needed, you can wipe the rest).
>
> To workaround the issue, just replace setnull with appropriate setter
for
> your root model element.
>
> Artem
>
> "Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de>
wrote
> in message news:epno3r$isa$1@utils.eclipse.org...
>> Okay I just found something related to this issue, but I don?t fully
>> understand. Where do I have to look for a failure?
>>
>> From: "Artem Tikhomirov":
>>
>> 'null' means generator failed to find a feature to set domainMetaElement
>> into DocumentRoot. Could you please check there's a stuctural feature
>> inside
>> DocumentRoot with domainMetaElement as eReferenceType?
>>
>> Artem
>>
>> "Jiefei" <majiefei@xxxxxxxxxxxxxxxxx> wrote in message
>> news:f89f488bf13cbbcf8d32a03d5e85f3a3$1@xxxxxxxxxxxxxxxxxx
>>> Hi,
>>> I mapped the DocumentRoot as the doaminMetaElement of the diagram.
>> But in
>>> the generated diagram code, the createInitialRoot(DocumentRoot model)
>>> method (in the FooDiagramEditorUtil class) has the implementation like:
>>>
>>> DocumentRoot docRoot = FooFactory.eINSTANCE.createDocumentRoot();
>>> docRoot.setnull(model);
>>> return docRoot;
>>>
>>> I dont understand why it generates "setnull". is it a bug?
>>>
>>> cheers,
>>>
>>> Jiefei
>>>
>>
>> Christian Danekas schrieb:
>>> Forgot to say that I?m using GMF 2.0 (19.01.2007 nightly).
>>>
>>> Christian Danekas schrieb:
>>>> Hello,
>>>>
>>>> I thought of using a xsd as the base for an editor. So I generated an
>>>> ecore and genmodel file from the xsd that we are using in our project.
>>>> Everything worked fine for me until code generation. In ecore models
>>>> derived from xsd models there is this artificial root element
>>>> Diagramroot. The generated code contains the following passage:
>>>>
>>>> /**
>>>> * @generated
>>>> */
>>>> private static EObject createInitialRoot(DocumentRoot model) {
>>>> DocumentRoot docRoot =
>>>> EditorFactory.eINSTANCE.createDocumentRoot();
>>>> docRoot.setnull(model);
>>>> return docRoot;
>>>> }
>>>>
>>>> The setnull method that is called does not exist and I have no idea
>>>> what is supposed to be done there. When I delete the line (thought
>>>> it?s worth a try) everything compiles fine, the editor application is
>>>> able to create a diagram but I am not able to place any nodes from the
>>>> palette onto the Canvas. The file containing the "semantic" model
>>>> contains nothing but this:
>>>>
>>>> <?xml version="1.0" encoding="UTF-8"?>
>>>> />
>>>>
>>>> Does anyone here have an idea whats going wrong there? The xsd is
>>>> valid so this can?t be the problem I think. Thank you in advance.
>>>>
>>>> Greetings,
>>>> Christian
>
>
>
>
|
|
|
| Re: Problem with GMF editor resulting from xsd [message #100644 is a reply to message #100617] |
Fri, 02 February 2007 10:45 |
| Eclipse User |
|
|
|
Hi,
> Mine looks exactly the same as the Mindmap model concerning the
> DocumentRoot Class. Instead of Map I have a reference to differently named
> class, which is also the diagram canvas. But there must be something wrong
> still.
Would be handy to get your ecore model (stripped down to DocumentRoot and
top level class) to take a look.
> I want my editor start the semantic model files like in the mindmap
> example where the DocumentRoot element can not be seen and the tree
> structure starts with <mindmap .
GMF's generated code does nothing special about the way elements get
persisted. That's up to EMF code. If you see DocumentRoot, check that
appropriate ResourceFactory gets registered.
> But when I changed the Extended metadata of my canvas element (the name
> started with lower case, I altered it to upper case) the behavior changed
> and the canvas plus DocumentRoot was created (before this change there was
> no canvas). I fail to see the logic behind this.
Me neither, though I'm not an XSD expert ;). I'm affraid, that's not GMF
question, .
"Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de> wrote
in message news:epvk4f$erj$1@utils.eclipse.org...
> Thank you, I managed to solve the problem by modifying the ecore model.
> Mine looks exactly the same as the Mindmap model concerning the
> DocumentRoot Class. Instead of Map I have a reference to differently named
> class, which is also the diagram canvas. But there must be something wrong
> still. When I generate all the code and launch the application I can
> create a new diagram. But sometimes there is no canvas element created at
> all and I can?t place my nodes on the Diagram, and other times there is
> first the DocumentRoot and the canvas I defined is below (in the semantic
> model file). I want my editor start the semantic model files like in the
> mindmap example where the DocumentRoot element can not be seen and the
> tree structure starts with <mindmap . I can?t find a difference between my
> models and the mindmap example. But when I changed the Extended metadata
> of my canvas element (the name started with lower case, I altered it to
> upper case) the behavior changed and the canvas plus DocumentRoot was
> created (before this change there was no canvas). I fail to see the logic
> behind this. Could you help me out again?
>
> Greetings,
> Christian
>
> Artem Tikhomirov schrieb:
> > <<untitled2.png>>
> > Please take a look at the screenshot attached. Map is
> domainMetaElement of
> > the MindMap domain model, and DocumentRoot.map feature (highlighted)
> is the
> > feature I was referring to below. You may want to check
> > org.eclipse.gmf.examples.mindmap (/model/mindmap.ecore) from GMF
> examples,
> > which is xsd-based model.
> >
> > There's some nasty logic detecting exact feature, perhaps, there's
> something
> > wrong with what's expected. Would be great if you provide us with
> copy of
> > .xsd file you are using (actually, definition of top xml elements
> only is
> > needed, you can wipe the rest).
> >
> > To workaround the issue, just replace setnull with appropriate setter
> for
> > your root model element.
> >
> > Artem
> >
> > "Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de>
> wrote
> > in message news:epno3r$isa$1@utils.eclipse.org...
> >> Okay I just found something related to this issue, but I don?t fully
> >> understand. Where do I have to look for a failure?
> >>
> >> From: "Artem Tikhomirov":
> >>
> >> 'null' means generator failed to find a feature to set
> >> domainMetaElement
> >> into DocumentRoot. Could you please check there's a stuctural feature
> >> inside
> >> DocumentRoot with domainMetaElement as eReferenceType?
> >>
> >> Artem
> >>
> >> "Jiefei" <majiefei@xxxxxxxxxxxxxxxxx> wrote in message
> >> news:f89f488bf13cbbcf8d32a03d5e85f3a3$1@xxxxxxxxxxxxxxxxxx
> >>> Hi,
> >>> I mapped the DocumentRoot as the doaminMetaElement of the diagram.
> >> But in
> >>> the generated diagram code, the createInitialRoot(DocumentRoot model)
> >>> method (in the FooDiagramEditorUtil class) has the implementation
> >>> like:
> >>>
> >>> DocumentRoot docRoot = FooFactory.eINSTANCE.createDocumentRoot();
> >>> docRoot.setnull(model);
> >>> return docRoot;
> >>>
> >>> I dont understand why it generates "setnull". is it a bug?
> >>>
> >>> cheers,
> >>>
> >>> Jiefei
> >>>
> >>
> >> Christian Danekas schrieb:
> >>> Forgot to say that I?m using GMF 2.0 (19.01.2007 nightly).
> >>>
> >>> Christian Danekas schrieb:
> >>>> Hello,
> >>>>
> >>>> I thought of using a xsd as the base for an editor. So I generated an
> >>>> ecore and genmodel file from the xsd that we are using in our
> >>>> project.
> >>>> Everything worked fine for me until code generation. In ecore models
> >>>> derived from xsd models there is this artificial root element
> >>>> Diagramroot. The generated code contains the following passage:
> >>>>
> >>>> /**
> >>>> * @generated
> >>>> */
> >>>> private static EObject createInitialRoot(DocumentRoot model) {
> >>>> DocumentRoot docRoot =
> >>>> EditorFactory.eINSTANCE.createDocumentRoot();
> >>>> docRoot.setnull(model);
> >>>> return docRoot;
> >>>> }
> >>>>
> >>>> The setnull method that is called does not exist and I have no idea
> >>>> what is supposed to be done there. When I delete the line (thought
> >>>> it?s worth a try) everything compiles fine, the editor application is
> >>>> able to create a diagram but I am not able to place any nodes from
> >>>> the
> >>>> palette onto the Canvas. The file containing the "semantic" model
> >>>> contains nothing but this:
> >>>>
> >>>> <?xml version="1.0" encoding="UTF-8"?>
> >>>> />
> >>>>
> >>>> Does anyone here have an idea whats going wrong there? The xsd is
> >>>> valid so this can?t be the problem I think. Thank you in advance.
> >>>>
> >>>> Greetings,
> >>>> Christian
> >
> >
> >
> >
|
|
|
| Re: Problem with GMF editor resulting from xsd [message #100668 is a reply to message #100644] |
Fri, 02 February 2007 10:54 |
| Eclipse User |
|
|
|
Originally posted by: merks.ca.ibm.com
This is a multi-part message in MIME format.
--------------030908070409080304020504
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Artem,
I'm watching this thread. I don't understand why it would be necessary
to tweak any of the Ecore model that's produced automatically from the
schema. This document shows how XML Schema annotations can be used to
affect names or other mappings:
http://www.eclipse.org/modeling/emf/docs/overviews/XMLSchema ToEcoreMapping.pdf
Your advice about ensuring that the right type of resource with the
right options (i.e., the generated XyzResourceFactoryImpl) is being used
is good.
If XSD expertise is required, I'll be happy to help out (as long as I
understand specifically what the issue is, which at this point I don't).
Artem Tikhomirov wrote:
> Hi,
>
>
>> Mine looks exactly the same as the Mindmap model concerning the
>> DocumentRoot Class. Instead of Map I have a reference to differently named
>> class, which is also the diagram canvas. But there must be something wrong
>> still.
>>
>
> Would be handy to get your ecore model (stripped down to DocumentRoot and
> top level class) to take a look.
>
>
>
>> I want my editor start the semantic model files like in the mindmap
>> example where the DocumentRoot element can not be seen and the tree
>> structure starts with <mindmap .
>>
>
> GMF's generated code does nothing special about the way elements get
> persisted. That's up to EMF code. If you see DocumentRoot, check that
> appropriate ResourceFactory gets registered.
>
>
>
>> But when I changed the Extended metadata of my canvas element (the name
>> started with lower case, I altered it to upper case) the behavior changed
>> and the canvas plus DocumentRoot was created (before this change there was
>> no canvas). I fail to see the logic behind this.
>>
>
> Me neither, though I'm not an XSD expert ;). I'm affraid, that's not GMF
> question, .
>
>
> "Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de> wrote
> in message news:epvk4f$erj$1@utils.eclipse.org...
>
>> Thank you, I managed to solve the problem by modifying the ecore model.
>> Mine looks exactly the same as the Mindmap model concerning the
>> DocumentRoot Class. Instead of Map I have a reference to differently named
>> class, which is also the diagram canvas. But there must be something wrong
>> still. When I generate all the code and launch the application I can
>> create a new diagram. But sometimes there is no canvas element created at
>> all and I can?t place my nodes on the Diagram, and other times there is
>> first the DocumentRoot and the canvas I defined is below (in the semantic
>> model file). I want my editor start the semantic model files like in the
>> mindmap example where the DocumentRoot element can not be seen and the
>> tree structure starts with <mindmap . I can?t find a difference between my
>> models and the mindmap example. But when I changed the Extended metadata
>> of my canvas element (the name started with lower case, I altered it to
>> upper case) the behavior changed and the canvas plus DocumentRoot was
>> created (before this change there was no canvas). I fail to see the logic
>> behind this. Could you help me out again?
>>
>> Greetings,
>> Christian
>>
>> Artem Tikhomirov schrieb:
>>
>>> <<untitled2.png>>
>>> Please take a look at the screenshot attached. Map is
>>>
>> domainMetaElement of
>>
>>> the MindMap domain model, and DocumentRoot.map feature (highlighted)
>>>
>> is the
>>
>>> feature I was referring to below. You may want to check
>>> org.eclipse.gmf.examples.mindmap (/model/mindmap.ecore) from GMF
>>>
>> examples,
>>
>>> which is xsd-based model.
>>>
>>> There's some nasty logic detecting exact feature, perhaps, there's
>>>
>> something
>>
>>> wrong with what's expected. Would be great if you provide us with
>>>
>> copy of
>>
>>> .xsd file you are using (actually, definition of top xml elements
>>>
>> only is
>>
>>> needed, you can wipe the rest).
>>>
>>> To workaround the issue, just replace setnull with appropriate setter
>>>
>> for
>>
>>> your root model element.
>>>
>>> Artem
>>>
>>> "Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de>
>>>
>> wrote
>>
>>> in message news:epno3r$isa$1@utils.eclipse.org...
>>>
>>>> Okay I just found something related to this issue, but I don?t fully
>>>> understand. Where do I have to look for a failure?
>>>>
>>>> From: "Artem Tikhomirov":
>>>>
>>>> 'null' means generator failed to find a feature to set
>>>> domainMetaElement
>>>> into DocumentRoot. Could you please check there's a stuctural feature
>>>> inside
>>>> DocumentRoot with domainMetaElement as eReferenceType?
>>>>
>>>> Artem
>>>>
>>>> "Jiefei" <majiefei@xxxxxxxxxxxxxxxxx> wrote in message
>>>> news:f89f488bf13cbbcf8d32a03d5e85f3a3$1@xxxxxxxxxxxxxxxxxx
>>>>
>>>>> Hi,
>>>>> I mapped the DocumentRoot as the doaminMetaElement of the diagram.
>>>>>
>>>> But in
>>>>
>>>>> the generated diagram code, the createInitialRoot(DocumentRoot model)
>>>>> method (in the FooDiagramEditorUtil class) has the implementation
>>>>> like:
>>>>>
>>>>> DocumentRoot docRoot = FooFactory.eINSTANCE.createDocumentRoot();
>>>>> docRoot.setnull(model);
>>>>> return docRoot;
>>>>>
>>>>> I dont understand why it generates "setnull". is it a bug?
>>>>>
>>>>> cheers,
>>>>>
>>>>> Jiefei
>>>>>
>>>>>
>>>> Christian Danekas schrieb:
>>>>
>>>>> Forgot to say that I?m using GMF 2.0 (19.01.2007 nightly).
>>>>>
>>>>> Christian Danekas schrieb:
>>>>>
>>>>>> Hello,
>>>>>>
>>>>>> I thought of using a xsd as the base for an editor. So I generated an
>>>>>> ecore and genmodel file from the xsd that we are using in our
>>>>>> project.
>>>>>> Everything worked fine for me until code generation. In ecore models
>>>>>> derived from xsd models there is this artificial root element
>>>>>> Diagramroot. The generated code contains the following passage:
>>>>>>
>>>>>> /**
>>>>>> * @generated
>>>>>> */
>>>>>> private static EObject createInitialRoot(DocumentRoot model) {
>>>>>> DocumentRoot docRoot =
>>>>>> EditorFactory.eINSTANCE.createDocumentRoot();
>>>>>> docRoot.setnull(model);
>>>>>> return docRoot;
>>>>>> }
>>>>>>
>>>>>> The setnull method that is called does not exist and I have no idea
>>>>>> what is supposed to be done there. When I delete the line (thought
>>>>>> it?s worth a try) everything compiles fine, the editor application is
>>>>>> able to create a diagram but I am not able to place any nodes from
>>>>>> the
>>>>>> palette onto the Canvas. The file containing the "semantic" model
>>>>>> contains nothing but this:
>>>>>>
>>>>>> <?xml version="1.0" encoding="UTF-8"?>
>>>>>> />
>>>>>>
>>>>>> Does anyone here have an idea whats going wrong there? The xsd is
>>>>>> valid so this can?t be the problem I think. Thank you in advance.
>>>>>>
>>>>>> Greetings,
>>>>>> Christian
>>>>>>
>>>
>>>
>>>
>
>
>
--------------030908070409080304020504
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Artem,<br>
<br>
I'm watching this thread. I don't understand why it would be necessary
to tweak any of the Ecore model that's produced automatically from the
schema. This document shows how XML Schema annotations can be used to
affect names or other mappings:<br>
<blockquote><a
href=" http://www.eclipse.org/modeling/emf/docs/overviews/XMLSchema ToEcoreMapping.pdf"> http://www.eclipse.org/modeling/emf/docs/overviews/XMLSchema ToEcoreMapping.pdf</a><br>
</blockquote>
Your advice about ensuring that the right type of resource with the
right options (i.e., the generated XyzResourceFactoryImpl) is being
used is good.<br>
<br>
If XSD expertise is required, I'll be happy to help out (as long as I
understand specifically what the issue is, which at this point I don't).<br>
<br>
<br>
Artem Tikhomirov wrote:
<blockquote cite="midepvmb7$jri$1@utils.eclipse.org" type="cite">
<pre wrap="">Hi,
</pre>
<blockquote type="cite">
<pre wrap="">Mine looks exactly the same as the Mindmap model concerning the
DocumentRoot Class. Instead of Map I have a reference to differently named
class, which is also the diagram canvas. But there must be something wrong
still.
</pre>
</blockquote>
<pre wrap=""><!---->
Would be handy to get your ecore model (stripped down to DocumentRoot and
top level class) to take a look.
</pre>
<blockquote type="cite">
<pre wrap="">I want my editor start the semantic model files like in the mindmap
example where the DocumentRoot element can not be seen and the tree
structure starts with <mindmap .
</pre>
</blockquote>
<pre wrap=""><!---->
GMF's generated code does nothing special about the way elements get
persisted. That's up to EMF code. If you see DocumentRoot, check that
appropriate ResourceFactory gets registered.
</pre>
<blockquote type="cite">
<pre wrap="">But when I changed the Extended metadata of my canvas element (the name
started with lower case, I altered it to upper case) the behavior changed
and the canvas plus DocumentRoot was created (before this change there was
no canvas). I fail to see the logic behind this.
</pre>
</blockquote>
<pre wrap=""><!---->
Me neither, though I'm not an XSD expert ;). I'm affraid, that's not GMF
question, .
"Christian Danekas" <a class="moz-txt-link-rfc2396E" href="mailto:christian.daenekas@informatik.uni-oldenburg.de"><christian.daenekas@informatik.uni-oldenburg.de></a> wrote
in message <a class="moz-txt-link-freetext" href="news:epvk4f$erj$1@utils.eclipse.org">news:epvk4f$erj$1@utils.eclipse.org</a>...
</pre>
<blockquote type="cite">
<pre wrap="">Thank you, I managed to solve the problem by modifying the ecore model.
Mine looks exactly the same as the Mindmap model concerning the
DocumentRoot Class. Instead of Map I have a reference to differently named
class, which is also the diagram canvas. But there must be something wrong
still. When I generate all the code and launch the application I can
create a new diagram. But sometimes there is no canvas element created at
all and I can?t place my nodes on the Diagram, and other times there is
first the DocumentRoot and the canvas I defined is below (in the semantic
model file). I want my editor start the semantic model files like in the
mindmap example where the DocumentRoot element can not be seen and the
tree structure starts with <mindmap . I can?t find a difference between my
models and the mindmap example. But when I changed the Extended metadata
of my canvas element (the name started with lower case, I altered it to
upper case) the behavior changed and the canvas plus DocumentRoot was
created (before this change there was no canvas). I fail to see the logic
behind this. Could you help me out again?
Greetings,
Christian
Artem Tikhomirov schrieb:
</pre>
<blockquote type="cite">
<pre wrap=""> <<untitled2.png>>
Please take a look at the screenshot attached. Map is
</pre>
</blockquote>
<pre wrap="">domainMetaElement of
</pre>
<blockquote type="cite">
<pre wrap="">the MindMap domain model, and DocumentRoot.map feature (highlighted)
</pre>
</blockquote>
<pre wrap="">is the
</pre>
<blockquote type="cite">
<pre wrap="">feature I was referring to below. You may want to check
org.eclipse.gmf.examples.mindmap (/model/mindmap.ecore) from GMF
</pre>
</blockquote>
<pre wrap="">examples,
</pre>
<blockquote type="cite">
<pre wrap="">which is xsd-based model.
There's some nasty logic detecting exact feature, perhaps, there's
</pre>
</blockquote>
<pre wrap="">something
</pre>
<blockquote type="cite">
<pre wrap="">wrong with what's expected. Would be great if you provide us with
</pre>
</blockquote>
<pre wrap="">copy of
</pre>
<blockquote type="cite">
<pre wrap="">.xsd file you are using (actually, definition of top xml elements
</pre>
</blockquote>
<pre wrap="">only is
</pre>
<blockquote type="cite">
<pre wrap="">needed, you can wipe the rest).
To workaround the issue, just replace setnull with appropriate setter
</pre>
</blockquote>
<pre wrap="">for
</pre>
<blockquote type="cite">
<pre wrap="">your root model element.
Artem
"Christian Danekas" <a class="moz-txt-link-rfc2396E" href="mailto:christian.daenekas@informatik.uni-oldenburg.de"><christian.daenekas@informatik.uni-oldenburg.de></a>
</pre>
</blockquote>
<pre wrap="">wrote
</pre>
<blockquote type="cite">
<pre wrap="">in message <a class="moz-txt-link-freetext" href="news:epno3r$isa$1@utils.eclipse.org">news:epno3r$isa$1@utils.eclipse.org</a>...
</pre>
<blockquote type="cite">
<pre wrap="">Okay I just found something related to this issue, but I don?t fully
understand. Where do I have to look for a failure?
From: "Artem Tikhomirov":
'null' means generator failed to find a feature to set
domainMetaElement
into DocumentRoot. Could you please check there's a stuctural feature
inside
DocumentRoot with domainMetaElement as eReferenceType?
Artem
"Jiefei" <majiefei@xxxxxxxxxxxxxxxxx> wrote in message
<a class="moz-txt-link-freetext" href="news:f89f488bf13cbbcf8d32a03d5e85f3a3$1@xxxxxxxxxxxxxxxxxx">news:f89f488bf13cbbcf8d32a03d5e85f3a3$1@xxxxxxxxxxxxxxxxxx</a>
</pre>
<blockquote type="cite">
<pre wrap="">Hi,
I mapped the DocumentRoot as the doaminMetaElement of the diagram.
</pre>
</blockquote>
<pre wrap="">But in
</pre>
<blockquote type="cite">
<pre wrap="">the generated diagram code, the createInitialRoot(DocumentRoot model)
method (in the FooDiagramEditorUtil class) has the implementation
like:
DocumentRoot docRoot = FooFactory.eINSTANCE.createDocumentRoot();
docRoot.setnull(model);
return docRoot;
I dont understand why it generates "setnull". is it a bug?
cheers,
Jiefei
</pre>
</blockquote>
<pre wrap="">Christian Danekas schrieb:
</pre>
<blockquote type="cite">
<pre wrap="">Forgot to say that I?m using GMF 2.0 (19.01.2007 nightly).
Christian Danekas schrieb:
</pre>
<blockquote type="cite">
<pre wrap="">Hello,
I thought of using a xsd as the base for an editor. So I generated an
ecore and genmodel file from the xsd that we are using in our
project.
Everything worked fine for me until code generation. In ecore models
derived from xsd models there is this artificial root element
Diagramroot. The generated code contains the following passage:
/**
* @generated
*/
private static EObject createInitialRoot(DocumentRoot model) {
DocumentRoot docRoot =
EditorFactory.eINSTANCE.createDocumentRoot();
docRoot.setnull(model);
return docRoot;
}
The setnull method that is called does not exist and I have no idea
what is supposed to be done there. When I delete the line (thought
it?s worth a try) everything compiles fine, the editor application is
able to create a diagram but I am not able to place any nodes from
the
palette onto the Canvas. The file containing the "semantic" model
contains nothing but this:
<?xml version="1.0" encoding="UTF-8"?>
/>
Does anyone here have an idea whats going wrong there? The xsd is
valid so this can?t be the problem I think. Thank you in advance.
Greetings,
Christian
</pre>
</blockquote>
</blockquote>
</blockquote>
<pre wrap="">
</pre>
</blockquote>
</blockquote>
<pre wrap=""><!---->
</pre>
</blockquote>
<br>
</body>
</html>
--------------030908070409080304020504--
|
|
|
| Re: Problem with GMF editor resulting from xsd [message #101361 is a reply to message #100644] |
Tue, 06 February 2007 05:02 |
| Eclipse User |
|
|
|
Originally posted by: christian.daenekas.informatik.uni-oldenburg.de
Thank you for your help Artem and Ed. Since the news server didn´t seem
to work the last 2 days I couldn´t write earlier. It´s really strange,
but after I changed the name attribute of my canvas element extended
metadata to "Metric" instead of "metric" and altered the ecore root
package name and the model file extension to "metric" and regenerated
the whole code it works and my model file starts with <metric> and the
canvas of the diagram is also linked to this element. In my desperation
I tried to synchronize the mindmap example and my editor as far as I
could (concerning the elements properties) and it worked.
Greetings,
Christian
Artem Tikhomirov schrieb:
> Hi,
>
>> Mine looks exactly the same as the Mindmap model concerning the
>> DocumentRoot Class. Instead of Map I have a reference to differently named
>> class, which is also the diagram canvas. But there must be something wrong
>> still.
>
> Would be handy to get your ecore model (stripped down to DocumentRoot and
> top level class) to take a look.
>
>
>> I want my editor start the semantic model files like in the mindmap
>> example where the DocumentRoot element can not be seen and the tree
>> structure starts with <mindmap .
>
> GMF's generated code does nothing special about the way elements get
> persisted. That's up to EMF code. If you see DocumentRoot, check that
> appropriate ResourceFactory gets registered.
>
>
>> But when I changed the Extended metadata of my canvas element (the name
>> started with lower case, I altered it to upper case) the behavior changed
>> and the canvas plus DocumentRoot was created (before this change there was
>> no canvas). I fail to see the logic behind this.
>
> Me neither, though I'm not an XSD expert ;). I'm affraid, that's not GMF
> question, .
>
>
> "Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de> wrote
> in message news:epvk4f$erj$1@utils.eclipse.org...
>> Thank you, I managed to solve the problem by modifying the ecore model.
>> Mine looks exactly the same as the Mindmap model concerning the
>> DocumentRoot Class. Instead of Map I have a reference to differently named
>> class, which is also the diagram canvas. But there must be something wrong
>> still. When I generate all the code and launch the application I can
>> create a new diagram. But sometimes there is no canvas element created at
>> all and I can?t place my nodes on the Diagram, and other times there is
>> first the DocumentRoot and the canvas I defined is below (in the semantic
>> model file). I want my editor start the semantic model files like in the
>> mindmap example where the DocumentRoot element can not be seen and the
>> tree structure starts with <mindmap . I can?t find a difference between my
>> models and the mindmap example. But when I changed the Extended metadata
>> of my canvas element (the name started with lower case, I altered it to
>> upper case) the behavior changed and the canvas plus DocumentRoot was
>> created (before this change there was no canvas). I fail to see the logic
>> behind this. Could you help me out again?
>>
>> Greetings,
>> Christian
>>
>> Artem Tikhomirov schrieb:
>>> <<untitled2.png>>
>>> Please take a look at the screenshot attached. Map is
>> domainMetaElement of
>>> the MindMap domain model, and DocumentRoot.map feature (highlighted)
>> is the
>>> feature I was referring to below. You may want to check
>>> org.eclipse.gmf.examples.mindmap (/model/mindmap.ecore) from GMF
>> examples,
>>> which is xsd-based model.
>>>
>>> There's some nasty logic detecting exact feature, perhaps, there's
>> something
>>> wrong with what's expected. Would be great if you provide us with
>> copy of
>>> .xsd file you are using (actually, definition of top xml elements
>> only is
>>> needed, you can wipe the rest).
>>>
>>> To workaround the issue, just replace setnull with appropriate setter
>> for
>>> your root model element.
>>>
>>> Artem
>>>
>>> "Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de>
>> wrote
>>> in message news:epno3r$isa$1@utils.eclipse.org...
>>>> Okay I just found something related to this issue, but I don?t fully
>>>> understand. Where do I have to look for a failure?
>>>>
>>>> From: "Artem Tikhomirov":
>>>>
>>>> 'null' means generator failed to find a feature to set
>>>> domainMetaElement
>>>> into DocumentRoot. Could you please check there's a stuctural feature
>>>> inside
>>>> DocumentRoot with domainMetaElement as eReferenceType?
>>>>
>>>> Artem
>>>>
>>>> "Jiefei" <majiefei@xxxxxxxxxxxxxxxxx> wrote in message
>>>> news:f89f488bf13cbbcf8d32a03d5e85f3a3$1@xxxxxxxxxxxxxxxxxx
>>>>> Hi,
>>>>> I mapped the DocumentRoot as the doaminMetaElement of the diagram.
>>>> But in
>>>>> the generated diagram code, the createInitialRoot(DocumentRoot model)
>>>>> method (in the FooDiagramEditorUtil class) has the implementation
>>>>> like:
>>>>>
>>>>> DocumentRoot docRoot = FooFactory.eINSTANCE.createDocumentRoot();
>>>>> docRoot.setnull(model);
>>>>> return docRoot;
>>>>>
>>>>> I dont understand why it generates "setnull". is it a bug?
>>>>>
>>>>> cheers,
>>>>>
>>>>> Jiefei
>>>>>
>>>> Christian Danekas schrieb:
>>>>> Forgot to say that I?m using GMF 2.0 (19.01.2007 nightly).
>>>>>
>>>>> Christian Danekas schrieb:
>>>>>> Hello,
>>>>>>
>>>>>> I thought of using a xsd as the base for an editor. So I generated an
>>>>>> ecore and genmodel file from the xsd that we are using in our
>>>>>> project.
>>>>>> Everything worked fine for me until code generation. In ecore models
>>>>>> derived from xsd models there is this artificial root element
>>>>>> Diagramroot. The generated code contains the following passage:
>>>>>>
>>>>>> /**
>>>>>> * @generated
>>>>>> */
>>>>>> private static EObject createInitialRoot(DocumentRoot model) {
>>>>>> DocumentRoot docRoot =
>>>>>> EditorFactory.eINSTANCE.createDocumentRoot();
>>>>>> docRoot.setnull(model);
>>>>>> return docRoot;
>>>>>> }
>>>>>>
>>>>>> The setnull method that is called does not exist and I have no idea
>>>>>> what is supposed to be done there. When I delete the line (thought
>>>>>> it?s worth a try) everything compiles fine, the editor application is
>>>>>> able to create a diagram but I am not able to place any nodes from
>>>>>> the
>>>>>> palette onto the Canvas. The file containing the "semantic" model
>>>>>> contains nothing but this:
>>>>>>
>>>>>> <?xml version="1.0" encoding="UTF-8"?>
>>>>>> />
>>>>>>
>>>>>> Does anyone here have an idea whats going wrong there? The xsd is
>>>>>> valid so this can?t be the problem I think. Thank you in advance.
>>>>>>
>>>>>> Greetings,
>>>>>> Christian
>>>
>>>
>>>
>
>
|
|
|
| Re: Problem with GMF editor resulting from xsd [message #101528 is a reply to message #101361] |
Tue, 06 February 2007 10:21 |
| Eclipse User |
|
|
|
Originally posted by: merks.ca.ibm.com
Christian,
This sounds likely to be related to what type of resource is being used
to save and load instances. When the option
XMLResource.OPTION_EXTENDED_META_DATA is used, the code use the original
XML names (as recorded in the extended meta data annotations) for
reading and writing; without that option, the code will use the actual
names of the classes and features as used in the Ecore model. The
handling of the root element is also different in each case. When using
extended meta data, the root element name of the instance must patch
exactly the XML name of an element-based feature in the DocumentRoot;
without extended metadata, the root element name is looked up as an
EClass in the EPackage corresponding to the element's namespace. Given
your comment about changing the file extension to metric, it's highly
likely that you are seeing behavior differences depending on which
resource factory is being used (and hence which options are being used
for load and save).
Christian Dänekas wrote:
> Thank you for your help Artem and Ed. Since the news server didn´t
> seem to work the last 2 days I couldn´t write earlier. It´s really
> strange, but after I changed the name attribute of my canvas element
> extended metadata to "Metric" instead of "metric" and altered the
> ecore root package name and the model file extension to "metric" and
> regenerated the whole code it works and my model file starts with
> <metric> and the canvas of the diagram is also linked to this element.
> In my desperation I tried to synchronize the mindmap example and my
> editor as far as I could (concerning the elements properties) and it
> worked.
>
> Greetings,
> Christian
>
> Artem Tikhomirov schrieb:
>> Hi,
>>
>>> Mine looks exactly the same as the Mindmap model concerning the
>>> DocumentRoot Class. Instead of Map I have a reference to differently
>>> named class, which is also the diagram canvas. But there must be
>>> something wrong still.
>>
>> Would be handy to get your ecore model (stripped down to DocumentRoot
>> and top level class) to take a look.
>>
>>
>>> I want my editor start the semantic model files like in the mindmap
>>> example where the DocumentRoot element can not be seen and the tree
>>> structure starts with <mindmap .
>>
>> GMF's generated code does nothing special about the way elements get
>> persisted. That's up to EMF code. If you see DocumentRoot, check that
>> appropriate ResourceFactory gets registered.
>>
>>
>>> But when I changed the Extended metadata of my canvas element (the
>>> name started with lower case, I altered it to upper case) the
>>> behavior changed and the canvas plus DocumentRoot was created
>>> (before this change there was no canvas). I fail to see the logic
>>> behind this.
>>
>> Me neither, though I'm not an XSD expert ;). I'm affraid, that's not
>> GMF question, .
>>
>>
>> "Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de>
>> wrote in message news:epvk4f$erj$1@utils.eclipse.org...
>>> Thank you, I managed to solve the problem by modifying the ecore
>>> model. Mine looks exactly the same as the Mindmap model concerning
>>> the DocumentRoot Class. Instead of Map I have a reference to
>>> differently named class, which is also the diagram canvas. But there
>>> must be something wrong still. When I generate all the code and
>>> launch the application I can create a new diagram. But sometimes
>>> there is no canvas element created at all and I can?t place my nodes
>>> on the Diagram, and other times there is first the DocumentRoot and
>>> the canvas I defined is below (in the semantic model file). I want
>>> my editor start the semantic model files like in the mindmap example
>>> where the DocumentRoot element can not be seen and the tree
>>> structure starts with <mindmap . I can?t find a difference between
>>> my models and the mindmap example. But when I changed the Extended
>>> metadata of my canvas element (the name started with lower case, I
>>> altered it to upper case) the behavior changed and the canvas plus
>>> DocumentRoot was created (before this change there was no canvas). I
>>> fail to see the logic behind this. Could you help me out again?
>>>
>>> Greetings,
>>> Christian
>>>
>>> Artem Tikhomirov schrieb:
>>>> <<untitled2.png>>
>>>> Please take a look at the screenshot attached. Map is
>>> domainMetaElement of
>>>> the MindMap domain model, and DocumentRoot.map feature (highlighted)
>>> is the
>>>> feature I was referring to below. You may want to check
>>>> org.eclipse.gmf.examples.mindmap (/model/mindmap.ecore) from GMF
>>> examples,
>>>> which is xsd-based model.
>>>>
>>>> There's some nasty logic detecting exact feature, perhaps, there's
>>> something
>>>> wrong with what's expected. Would be great if you provide us with
>>> copy of
>>>> .xsd file you are using (actually, definition of top xml elements
>>> only is
>>>> needed, you can wipe the rest).
>>>>
>>>> To workaround the issue, just replace setnull with appropriate setter
>>> for
>>>> your root model element.
>>>>
>>>> Artem
>>>>
>>>> "Christian Danekas" <christian.daenekas@informatik.uni-oldenburg.de>
>>> wrote
>>>> in message news:epno3r$isa$1@utils.eclipse.org...
>>>>> Okay I just found something related to this issue, but I don?t fully
>>>>> understand. Where do I have to look for a failure?
>>>>>
>>>>> From: "Artem Tikhomirov":
>>>>>
>>>>> 'null' means generator failed to find a feature to set
>>>>> domainMetaElement
>>>>> into DocumentRoot. Could you please check there's a stuctural feature
>>>>> inside
>>>>> DocumentRoot with domainMetaElement as eReferenceType?
>>>>>
>>>>> Artem
>>>>>
>>>>> "Jiefei" <majiefei@xxxxxxxxxxxxxxxxx> wrote in message
>>>>> news:f89f488bf13cbbcf8d32a03d5e85f3a3$1@xxxxxxxxxxxxxxxxxx
>>>>>> Hi,
>>>>>> I mapped the DocumentRoot as the doaminMetaElement of the diagram.
>>>>> But in
>>>>>> the generated diagram code, the createInitialRoot(DocumentRoot
>>>>>> model)
>>>>>> method (in the FooDiagramEditorUtil class) has the implementation
>>>>>> like:
>>>>>>
>>>>>> DocumentRoot docRoot = FooFactory.eINSTANCE.createDocumentRoot();
>>>>>> docRoot.setnull(model);
>>>>>> return docRoot;
>>>>>>
>>>>>> I dont understand why it generates "setnull". is it a bug?
>>>>>>
>>>>>> cheers,
>>>>>>
>>>>>> Jiefei
>>>>>>
>>>>> Christian Danekas schrieb:
>>>>>> Forgot to say that I?m using GMF 2.0 (19.01.2007 nightly).
>>>>>>
>>>>>> Christian Danekas schrieb:
>>>>>>> Hello,
>>>>>>>
>>>>>>> I thought of using a xsd as the base for an editor. So I
>>>>>>> generated an
>>>>>>> ecore and genmodel file from the xsd that we are using in our
>>>>>>> project.
>>>>>>> Everything worked fine for me until code generation. In ecore
>>>>>>> models
>>>>>>> derived from xsd models there is this artificial root element
>>>>>>> Diagramroot. The generated code contains the following passage:
>>>>>>>
>>>>>>> /**
>>>>>>> * @generated
>>>>>>> */
>>>>>>> private static EObject createInitialRoot(DocumentRoot model) {
>>>>>>> DocumentRoot docRoot =
>>>>>>> EditorFactory.eINSTANCE.createDocumentRoot();
>>>>>>> docRoot.setnull(model);
>>>>>>> return docRoot;
>>>>>>> }
>>>>>>>
>>>>>>> The setnull method that is called does not exist and I have no idea
>>>>>>> what is supposed to be done there. When I delete the line (thought
>>>>>>> it?s worth a try) everything compiles fine, the editor
>>>>>>> application is
>>>>>>> able to create a diagram but I am not able to place any nodes
>>>>>>> from the
>>>>>>> palette onto the Canvas. The file containing the "semantic" model
>>>>>>> contains nothing but this:
>>>>>>>
>>>>>>> <?xml version="1.0" encoding="UTF-8"?>
>>>>>>> />
>>>>>>>
>>>>>>> Does anyone here have an idea whats going wrong there? The xsd is
>>>>>>> valid so this can?t be the problem I think. Thank you in advance.
>>>>>>>
>>>>>>> Greetings,
>>>>>>> Christian
>>>>
>>>>
>>>>
>>
>>
|
|
|
Goto Forum:
Current Time: Fri May 09 18:21:43 EDT 2025
Powered by FUDForum. Page generated in 0.04162 seconds |
Home