| EMF Compare of combined models, comparing packages at new different levels [message #956836] |
Wed, 24 October 2012 17:00  |
Eclipse User |
|
|
|
I previously had a structure like so;
org.whatever.mypackage.ecore
->EPackage#org.whatever.mypackage
org.whatever.mypackage.mysubpacakage1.ecore
->EPackage#org.whatever.mypackage.mysubpacakage1
org.whatever.mypackage.mysubpacakage2.ecore
->EPackage#org.whatever.mypackage.mysubpacakage2
where each ecore just had its own package, and subpackages were in separate ecore files. However we moved this to a single ecore where appropriate like so;
org.whatever.mypackage.ecore
->EPackage#org.whatever.mypackage
->eSubpackages#org.whatever.mypackage.mysubpacakage1
->eSubpackages#org.whatever.mypackage.mysubpacakage2
Is there any way to explicitly compare across different levels?
I am thinking that 1) I would either highlight the packages I want to compare in the ecore model editor and select compare with each other, or that 2) I would reconstruct each hierarchy by loading the resources, and compare that. But I don't see any options.
Any Suggestions?
Thanks,
Tom
|
|
|
| Re: EMF Compare of combined models, comparing packages at new different levels [message #957483 is a reply to message #956836] |
Thu, 25 October 2012 04:56  |
| Eclipse User |
|
|
|
Tom,
First things first, which version of EMF Compare are you using?
EMF Compare 2.0 (and 1.3.1) will both load the whole logical model when comparing. This means that when you try and compare "mypackage.ecore" with another model, EMF Compare will load "mypackage.ecore" and all of the other resources it reference, in this case "mysubpacakage1.ecore" and "mysubpacakage2.ecore". In short, if you simply compare your two versions of "mypackage.ecore", EMF Compare should just work. Note that this assumes that "mypackage" does indeed reference the two sub-packages (i.e. when you load "mypackage" in an ecore editor and right-click > validate it, you have the two other resources loaded in that editor : they are "fragments" of mypackage).
2.0 will see (at least) two differences that correspond to the "unfragmenting" of your model (i.e. the move of subpackage1 and subpackage2 into mypackage). Version 1.* had a limitation that prevented us from detecting that.
[edit : you might be interested in my rant about the Logical models in EMF Compare 2...]
Laurent Goubet
Obeo
[Updated on: Thu, 25 October 2012 04:57] by Moderator |
|
|
Powered by
FUDForum. Page generated in 0.39630 seconds
Home