[news.eclipse.technology.epf] Re: EPF XMI file parsing problem

[Ed Merks <merks@xxxxxxxxxx>]

Sebastien,

When you get your best friend, the debugger, to help, can you tell 
what's happening that's at the line that's causing a null pointer 
exception to occur.  The fact that you have UI dependencies in your set 
of dependencies makes me suspicious that you either have too many 
dependencies or the dependencies won't work standalone.  I don't have 
EPF source code sitting around so at best I can provide advice based on 
the information your make available about what the code is doing.  So I 
can see that creating an instance of  ActivityDetailDiagram during 
deserialization is causing some adapter factory to be initialized:

   at 
org.eclipse.epf.library.edit.internal.TngAdapterFactoryImpl.getProcessComposedAdapterFactory(Unknown 
Source)
   at org.eclipse.epf.library.edit.TngAdapterFactory.<clinit>(Unknown 
Source)
   at 
org.eclipse.epf.diagram.model.impl.ActivityDetailDiagramImpl.<clinit>(Unknown 
Source)

It's already gone of into some "edit" code, which if it's based on 
org.eclipse.emf.edit should still function standalone.  But it's going 
off to ask something of some Providers thing and that getting the null 
pointer exception for mystery reasons since I don't have the source:

Caused by: java.lang.NullPointerException
   at org.eclipse.epf.library.edit.Providers.<clinit>(Unknown Source)

What's this line doing?  And the bigger question, why is "edit" code 
needed purely when loading a model.  I'm not saying there's anything 
wrong with that---the Mapping framework, for example, depends on edit 
code too---I'm just curious....


Sebastien BALARD wrote:
> Ed Merks wrote:
> > Sebastien,
> >
> > I guess it's more of an EPF question since the null pointer exception 
> > is here.
> >
> >     org.eclipse.epf.library.edit.Providers
> >
> > It might well be that the model doesn't work standalone
>
> Can anyone confirm that ?
>
> > or that you've missed some other important registrations that would 
> > normally happen...
>
> or, in this case, can anyone explain to me what I've missed ?
>
> Regards,
> Sebastien