| using UUID for/in HREF [message #426055] |
Mon, 15 December 2008 08:10  |
Eclipse User |
|
|
|
Hello everyone,
I have a semantic model which is split into several resources.
The problem is that the uri references are problematic with lazy loading
resources and keeping the references consistent.
I have now changed the (de-)serialization to use UUIDs (by using the
GMFResourceFactory).
Now I have a unique identifier for each Object in a resource, but the
references are still stored with "path_to_resource#//Containment.n". Is it
possible to use the resource path and the uuid instead?
I.e., I would like to have something like "path_to_resource#UUID" for
resolving the object reference?
Thanks in advance!
Soeren
|
|
|
| Re: using UUID for/in HREF [message #426056 is a reply to message #426055] |
Mon, 15 December 2008 08:14  |
| Eclipse User |
|
|
|
Soeren,
Comments below.
Soeren Kemmann wrote:
> Hello everyone,
>
> I have a semantic model which is split into several resources.
> The problem is that the uri references are problematic with lazy loading
> resources and keeping the references consistent.
> I have now changed the (de-)serialization to use UUIDs (by using the
> GMFResourceFactory).
> Now I have a unique identifier for each Object in a resource, but the
> references are still stored with "path_to_resource#//Containment.n". Is it
> possible to use the resource path and the uuid instead?
>
That's normally what ought to happen.
> I.e., I would like to have something like "path_to_resource#UUID" for
> resolving the object reference?
>
If you load the referencing resource, ensure that all proxies are
resolved, e.g., EcoreUtil.resolveAll or invoking Validate... in the
editor, and then save, it should use the new style references. Note
that you could also defines eKeys for your containment references so
that you'd get references of the form @feature[key='value']. UUIDs are
not cheap. They have an overhead of roughly 175 bytes per object...
> Thanks in advance!
>
> Soeren
>
>
>
|
|
|
Powered by
FUDForum. Page generated in 0.11174 seconds
Home