|
Re: Group Differences [message #1765942 is a reply to message #1765873] |
Wed, 14 June 2017 14:27 |
|
Hi Simon
I'm not sure I completely understand the question since, in the example you describe, there should only be 2 difference:
Quote:
eType EString [java.lang.String] has been set
Foo : EString has been added to eStructuralFeatures
The other one I don't know the origin of?
That being said, the UI uses filters to remove the first of these two, namely the "cascading differences" filter.
It sounds like you're not using the EMF Compare UI; are you trying to re-code one?
If you're using the default EMF Compare UI, you should have a look at the filters, they're extensible so you can add filters for your specific use cases if the cascading one doesn't meet your needs.
If you're trying to re-code an UI for the comparison, you might want to look at how we implemented filtering on the tree viewers, specifically, you want to look at org.eclipse.emf.compare.rcp.ui.internal.structuremergeviewer.filters.StructureMergeViewerFilter and its uses.
This is kind of a scattershot answer, could you describe your use case a little more so that I might understand better what you're trying to achieve?
Laurent Goubet
Obeo
|
|
|
|
Re: Group Differences [message #1766501 is a reply to message #1766411] |
Thu, 22 June 2017 09:24 |
|
Hi Simon,
Yes if you're displaying the differences without using the viewers provided by EMF Compare you will likely have to recode your own filter. The main class used for this is "ViewerFilter" from swt that can be applied to most viewers.
Do note that we are not happy with how the CascadingDifferencesFilter works (mainly because the filter, an UI mechanism, changes the way the merge, a core functionality, will handle differences). The mechanics underlying this are likely to change in the next version so that the "sub" differences are somehow recognizable through a link between them (for now, we don't really know how to represent that. Refinement maybe?). Please let us know of any input you might have on that and how you think this would be better described in the comparison model.
Laurent Goubet
Obeo
|
|
|
Powered by
FUDForum. Page generated in 0.25154 seconds