Bugzilla – Full Text Bug Listing
| Summary: | Fields of the model facet in the gmfgen are not editable in the gmfgen editor | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Modeling] GMF-Runtime | Reporter: | Patrick Tessier <Patrick.Tessier> | ||||||||
| Component: | Notation | Assignee: | Lidija Grahek <lgrahek> | ||||||||
| Status: | RESOLVED FIXED | QA Contact: | |||||||||
| Severity: | critical | ||||||||||
| Priority: | P3 | CC: | aboyko, ahunter.eclipse, Ed.Merks, lgrahek, mariot.chauvin | ||||||||
| Version: | 2.1 | Flags: | Ed.Merks:
pmc_approved+
mariot.chauvin: review? (ahunter.eclipse) |
||||||||
| Target Milestone: | --- | ||||||||||
| Hardware: | All | ||||||||||
| OS: | All | ||||||||||
| Whiteboard: | |||||||||||
| Attachments: |
|
||||||||||
Hi Patrick, This error log generally indicates that the referenced eObject could not be found, it does not resolve. Could you check that the eObject you reference are accessible (have a look to the URI) ? Could you provide the gmfgen file, or at least the URIs which do not resolve ? Thanks, Mariot The gmf gnen model that can be found : http://dev.eclipse.org/svnroot/modeling/org.eclipse.mdt.papyrus/trunk/plugins/uml/org.eclipse.papyrus.diagram.clazz/model/classdiagram.gmfgen in the directory http://dev.eclipse.org/svnroot/modeling/org.eclipse.mdt.papyrus/trunk/plugins/uml/org.eclipse.papyrus.diagram.clazz/model/ to open it you need to have the following plugin http://dev.eclipse.org/svnroot/modeling/org.eclipse.mdt.papyrus/trunk/plugins/developer/org.eclipse.papyrus.gmfgenextension , because we use the extension of the gmf gen. It is strange because, when i open the file all runs, I can generate the code. the bug comes from when I try to edit the gmfgen file, when the editor looks for all possible values in order to create the combo box. I have followed with the debugger, it seems that it does not find a super class so the array contains a value null. (In reply to comment #2) > The gmf gnen model that can be found : > http://dev.eclipse.org/svnroot/modeling/org.eclipse.mdt.papyrus/trunk/plugins/uml/org.eclipse.papyrus.diagram.clazz/model/classdiagram.gmfgen > in the directory > http://dev.eclipse.org/svnroot/modeling/org.eclipse.mdt.papyrus/trunk/plugins/uml/org.eclipse.papyrus.diagram.clazz/model/ > > to open it you need to have the following plugin > http://dev.eclipse.org/svnroot/modeling/org.eclipse.mdt.papyrus/trunk/plugins/developer/org.eclipse.papyrus.gmfgenextension > , because we use the extension of the gmf gen. > > It is strange because, when i open the file all runs, I can generate the code. > the bug comes from when I try to edit the gmfgen file, when the editor looks > for all possible values in order to create the combo box. > I have followed with the debugger, it seems that it does not find a super class > so the array contains a value null. Patrick could you debug on your side, and say which eClasses are in parameters of collectGenClasses which returns a array with a null value ? At each time, this is the ecore class RoutingStyle that bugs
at debugger it contains as eAllSupertype:
- Style
- RoundedCornersStyle
At this line
GenClassImpl.getLabelFeatureCandidates() line: 2043
public List<GenFeature> getLabelFeatureCandidates()
{
return
collectGenFeatures
(getAllBaseGenClasses(),
getGenFeatures(),
new GenFeatureFilter()
{
public boolean accept(GenFeature genFeature)
{
return !genFeature.isReferenceType() && !genFeature.isListType() && !genFeature.isMapType();
}
});
}
getAllBaseGenClasses() seems to collect all esuperclass
see code in GenClassImpl:
public List<GenClass> getAllBaseGenClasses()
{
return collectGenClasses(getEcoreClass().getEAllSuperTypes(), null);
}
Normally we have to obtain this two superclasse: Style and RoundedCornersStyle
but we obtain a list with the size =2 in which RoundedCornersStyle is absent.
We have only style in the list but the size equal 2.
(In reply to comment #4) > At each time, this is the ecore class RoutingStyle that bugs > at debugger it contains as eAllSupertype: > - Style > - RoundedCornersStyle > > At this line > GenClassImpl.getLabelFeatureCandidates() line: 2043 > > public List<GenFeature> getLabelFeatureCandidates() > { > return > collectGenFeatures > (getAllBaseGenClasses(), > getGenFeatures(), > new GenFeatureFilter() > { > public boolean accept(GenFeature genFeature) > { > return !genFeature.isReferenceType() && !genFeature.isListType() > && !genFeature.isMapType(); > } > }); > } > getAllBaseGenClasses() seems to collect all esuperclass > see code in GenClassImpl: > > public List<GenClass> getAllBaseGenClasses() > { > return collectGenClasses(getEcoreClass().getEAllSuperTypes(), null); > } > > Normally we have to obtain this two superclasse: Style and RoundedCornersStyle > but we obtain a list with the size =2 in which RoundedCornersStyle is absent. > > We have only style in the list but the size equal 2. Did you have an old version of GMF installed in your platform ? I have checked that "NotationPackage.eINSTANCE.getRoundedCornersStyle();" in RC3 does not return null. I have downloaded the RC3 version an my version GMF is 2.3 20100508. I have also to remove the extension of the gmfgen, but the bug remains. Have you succeed to edit a gmfgen? Several colleagueq have the same problem. I was able to reproduce. The notation.genmodel is not up to date, it does not contain the genClass for RoundedCornerStyle ! I assigned myself. Created attachment 171506 [details]
patch for notation.genmodel file
According to Modeling ramp down policy, I am asking a review and pmc approval to commit this fix. Hi Lidija, It looks like we must have forgotten to commit the notation.genmodel changes with Bug 309221 . The notation.ecore file has your changes, but not the notation.genmodel . I refreshed and more than just Mariot's change is missing. Lidija, can you have a look? Are we saying that the GMF Tooling requires the notation.genmodel ? I never realized this. The notation.genmodel is not deplyed at runtime. So the patch to notation.genmodel does not actually fix any notation code. Mariot, what is missing here? Created attachment 171532 [details]
Full notation.genmodel patch
I attached the full patch for notation.genmodel in case it needs to be committed.
(In reply to comment #11) > The notation.genmodel is not deplyed at runtime. So the patch to > notation.genmodel does not actually fix any notation code. > > Mariot, what is missing here? The notation.genmodel is used by EMF which read it with this uri "platform:/plugin/org.eclipse.gmf.runtime.notation/model/notation.genmodel" due to the fact that gmfgen models reference it. (In reply to comment #13) > (In reply to comment #11) > > The notation.genmodel is not deplyed at runtime. So the patch to > > notation.genmodel does not actually fix any notation code. > > > > Mariot, what is missing here? > > The notation.genmodel is used by EMF which read it with this uri > "platform:/plugin/org.eclipse.gmf.runtime.notation/model/notation.genmodel" due > to the fact that gmfgen models reference it. Anthony could I commit the full notation.genmodel patch from Lidija ? (In reply to comment #11) > The notation.genmodel is not deplyed at runtime. So the patch to > notation.genmodel does not actually fix any notation code. OK, and the notation.genmodel is deployed at runtime as the models folder is delivered in the org.eclipse.gmf.runtime.notation bundle. Unfortunately, we are going to have to put this in 1.4.1 SR1. We can update the interim builds, we can point them at the update site. Or is this a GMF Tooling will not work for Helios unless we fix this? Hi Lidija, This needs to be committed to R1_4_maintenance and HEAD Committed to both R1_4_maintenance and HEAD. |
Created attachment 168932 [details] this file contains the error log + the list of used features Hi, I'm commeter in the papyrus project when I open a gmfgen with the gmfgen editor. Fields that reference eobjects can not be edited. for example: I can not succeed to edit fields in the model facet or fields in the extension gmfgen. The attachement is the log.