atom feed4 messages in org.eclipse.modeling-pmcRe: [modeling-pmc] non respect of the...
FromSent OnAttachments
Grégoire DupéApr 13, 2012 1:19 am 
Ed MerksApr 15, 2012 3:30 pm 
Grégoire DupéApr 16, 2012 1:45 am 
Ed MerksApr 16, 2012 3:05 am 
Subject:Re: [modeling-pmc] non respect of the API freeze by EMF Facet
From:Grégoire Dupé (
Date:Apr 16, 2012 1:45:23 am


Do you know who your primary consumers are at Eclipse?

As far as I know, the primary consumer is Papyrus. I regularly inform them (via
Vincent Lorenzo) of changes in EMF Facet.

Regards, Grégoire

----- Mail Original ----- De: "Ed Merks" <> À: "PMC members mailing list" <> Envoyé: Lundi 16 Avril 2012 00h31:25 GMT +01:00 Amsterdam / Berlin / Berne /
Rome / Stockholm / Vienne Objet: Re: [modeling-pmc] non respect of the API freeze by EMF Facet



Comments below.

On 13/04/2012 10:19 AM, Grégoire Dupé wrote:


Regarding the development plan, the development of EMF Facet 0.2 is relay late.
Main enhancements required by Papyrus have not been achieved before M6. We are
not so far to finish, but we need to commit modifications which would not
respect the M6 API freeze:

1) The plug-in org.eclipse.emf.facet.aggregate.ui has been released (in 0.2 M6)
but it is totally useless and it doesn’t contain any public API. It would be
really useful to remove it in the M7 release but another project may have a
dependency to the plug-in 'org.eclipse.emf.facet.aggregate.ui'. Do you know who your primary consumers are at Eclipse?

Cf. Bug 376574 - [Aggregate] Removing of the plug-in

2) The meta-model contained in aggregate-0.2.0.ecore contains too many useless
attributes and references. It would be helpful to remove them rather than have
to deal with them (for the users and the maintainers).

Cf. Bug 376575 - [Aggregate] aggregate-0.2.0.ecore cleanning

3) In the facet meta-model, the eReference Facet::extendedFacet hasn’t the right
multiplicity (should be -1 instead of 1).

Cf. Bug 376576 - [EFacet] Change the multiplicity of Facet::extendedFacet

4) The interface
org.eclipse.emf.facet.efacet.metamodel.v0_2_0.efacet.query.QueryPackage is not
marked with the tags ‘@noimplement’ and ‘@noextend’. This avoids adding new
eClasses without to break the API. It would be really useful to add them before
the M7 release.

Cf. Bug 376577 - [EFacet] '@noimplement' and '@noextend' in QueryPackage

To my mind it would better for the community to do the changes now (before M7)
rather to get ugly features for a few years. Furthermore, EMF Facet is in the
incubation state. I think we should benefit from this state to provide an as
clean as possible product.

I agree. I think it's generally best to fix glaring problems sooner rather than later.

That is why I ask the PMC to allow the project 'EMF Facet' to not respect the
API freeze.

In the end, you know your community best and you know what's best for your technology, so I'm inclined to approve whatever you think is best in this regard. Thanks for asking for permission.

Regards, Grégoire Dupé