| EObjectEList isUnique and EReferences that are not unique [message #417010] |
Fri, 22 February 2008 12:29  |
Eclipse User |
|
|
|
I see from https://bugs.eclipse.org/bugs/show_bug.cgi?id=89325
that its a known "problem" that EMF does not generate code for EReferences
which are NOT unique which deal with the non uniqueness.
I can't see people modeling non unique references very often, but they
could still hand me models which have this set, though no one has. Do you
have any intention, beyond the workaround, to do any work in this area?
Don't do it because I asked because I can't think of any real examples
that do it...
Thanks
|
|
|
| Re: EObjectEList isUnique and EReferences that are not unique [message #417016 is a reply to message #417010] |
Fri, 22 February 2008 14:07  |
| Eclipse User |
|
|
|
JF,
We always have good intentions, and in a sense fixing this would be
trivial. But the downstream implications are a little overwhelming. So
much code assumes that an object will occur at most once in the list.
It does seem to be extremely rare. The last time someone wanted this,
they needed to define like a "signature", i.e., a list of types. But in
the end they ended up with a parameter list and the problem went way.
It would take many weeks of careful investigation to track down the
implications of the change, so it won't happen this release for sure.
JF wrote:
> I see from https://bugs.eclipse.org/bugs/show_bug.cgi?id=89325
> that its a known "problem" that EMF does not generate code for
> EReferences which are NOT unique which deal with the non uniqueness.
>
> I can't see people modeling non unique references very often, but they
> could still hand me models which have this set, though no one has. Do
> you have any intention, beyond the workaround, to do any work in this
> area? Don't do it because I asked because I can't think of any real
> examples that do it...
>
> Thanks
>
|
|
|
|
Powered by
FUDForum. Page generated in 0.03582 seconds
Home