[EMF Compare] Is it possible to ignore the element order? [message #632564] |
Wed, 13 October 2010 09:13  |
Eclipse User |
|
|
|
Hi,
I'm trying to compare a refactored model to a specified target model
using EMF Compare. This is working fine and as expected but I'd like to
give the user an option to ignore re-ordered elements. Because the user
has to manually create a target model this might be more convenient than
making sure the elements have the same order as in the source model.
However, it seems to me there is no way to make EMF Compare ignore the
order.
Am I just overlooking something or is this not possible? I noticed the
papers mention a change of order as a basic possible difference between
models but in certain cases, semantically, reordering is not a
difference at all.
thanks
Jan
|
|
|
|
|
Re: [EMF Compare] Is it possible to ignore the element order? [message #633084 is a reply to message #632564] |
Fri, 15 October 2010 06:05  |
Eclipse User |
|
|
|
Hi Jan,
I had a similar issue and maybe this is or interest to you.
The comparisons I have contain a lot of order changes but I am not interested in them.
I'm calling EMF Compare programmatically, so I can access the computed diffmodel.
Before opening the compare UI for the user, I remove all order changes:
org.eclipse.emf.compare.mpatch.util.MPatchUtil.removeDiffEle mentOfType(snapshot.getDiff(), Collections.singleton(DiffPackage.Literals.REFERENCE_ORDER_C HANGE));
MPatchUtil is available in the CVS repository of EMF Compare, or I can post the method here if you like.
Hope it helps
Patrick
On 13-10-2010 15:13, Jan Baart wrote:
> Hi,
>
> I'm trying to compare a refactored model to a specified target model using EMF Compare. This is working fine and as expected but I'd like to give the user an option to ignore re-ordered elements. Because the user has to manually create a target model this might be more convenient than making sure
> the elements have the same order as in the source model. However, it seems to me there is no way to make EMF Compare ignore the order.
>
> Am I just overlooking something or is this not possible? I noticed the papers mention a change of order as a basic possible difference between models but in certain cases, semantically, reordering is not a difference at all.
>
> thanks
>
> Jan
|
|
|
Powered by
FUDForum. Page generated in 0.04115 seconds