[news.eclipse.technology.emft] Re: [EMF Compare] Understanding EMF Compa

[laurent Goubet <laurent.goubet@xxxxxxx>]

Hi Timothy,

In the case described, there is no way to discriminate between two 
elements other than through their name attribute. Between "Real" and 
"Double" Iéd have thought EMF Compare wouldn't be able to match those 
two elements. Do you have IDs/XMI IDs in your model? If yes, take a look 
at MatchOptions.OPTION_IGNORE_(XMI_)ID or if you're comparing through 
the GUI nad not code, the option window => preferences => EMF Compare => 
"ignore (xmi) id while matching".

IDs or not, since EMF Compare matched your two elements (you've got an 
UpdateAttribute instead of two "ModelElementChange" differences); it's 
only natural that merging the difference makes the second model into a 
copy of the first. It is possible to manually create the diff, but would 
 be better to try and determine how your two PrimitiveTypes can be 
similar enough for EMF Compare to match them.

Laurent Goubet
Obeo

Timothy Marc a écrit :
> Hello all Comparer,
>
> i'm a newbie to EMF Compare and i want to understand this great stuff. I 
> have a model with four elements
>
> Element:Root
> PrimitiveType{name=Integer}
> PrimitiveType{name=Real}
> PrimitiveType{name=Boolean}
> PrimitiveType{name=String}
>
> Than, i create another similiar model like this
>
> Element:Root
> PrimitiveType{name=Integer}
> PrimitiveType{name=Double}
> PrimitiveType{name=Boolean}
> PrimitiveType{name=String}
>
> Only the Real value changed to Double. After merging these two models, 
> the resulting models is exactly the second model. I've introspected the 
> DiffModel and a Update difference was detected.
>
> What i want the merge algorithm to do is to copy the new double 
> primitive type into the existing model, so that five primitive types are 
> contained in the root instead of four. So, for me a general questions 
> has raised:
>
> Is it possible to build the diff of these two models and than just copy 
> the missing elements into the left model? I didn't want to get the 
> models merged entiredly in the end.
>
> And how did EMF Compare indicated that two instances are "equal"?
>
> Thx
> Timothy

begin:vcard
fn:Laurent Goubet
n:Goubet;Laurent
org:<a href="http://www.obeo.fr/";>Obeo</a>
email;internet:laurent.goubet@xxxxxxx
url:http://www.obeo.fr
version:2.1
end:vcard