[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
RE: [m2m-dev] A home for QVT.mdl
|
Hi Ed,
QVTOML plugins have the following naming scheme,
'org.eclipse.m2m.qvt.oml.*'.
Would that be possible to adapt your models for operational QVT to this?
Regards,
/Radek
> -----Original Message-----
> From: m2m-dev-bounces@xxxxxxxxxxx
> [mailto:m2m-dev-bounces@xxxxxxxxxxx] On Behalf Of Ed Willink
> Sent: Saturday, May 31, 2008 9:01 AM
> To: m2m-dev@xxxxxxxxxxx
> Subject: [m2m-dev] A home for QVT.mdl
>
> Hi All
>
> The OMG QVT diagrams and models are all notionally generated
> from a QVT_1.0.mdl.
> Unfortunately there are numerous errors at every stage of
> often manual transcription.
> An OMG issue on each error in each of the QVT 20070708 Ecore
> models will be submitted shortly.
>
> The models recently contributed to QVT Declarative
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=234220,
> http://dev.eclipse.org/ipzilla/show_bug.cgi?id=2342
> are auto-generated from a bug-fixed QVT.mdl with a *.cat per
> package and a variant of org.eclipse.emf.importer.rose that
> correctly handles "{ordered}". Unnavigable opposite role
> names are modelled as EMOF Comments/Ecore annotations (see
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=229998) and the
> org.eclipse.ocl.AbstractEnvironment.findNonNavigableAssociatio
> nEnds() stub.
>
> [The opposite role names will be in an update to the contribution.
> Development of JUnit tests to generate the OMG issues showed
> that I'd overlooked the opposites.]
>
> [Similar models are ready for contribution to QVTOML. I just
> need confirmation of the required package naming. Names
> currently assume e.g. org.eclipse.qvtoml.ecore.imperativeocl]
>
> The never-CVSed UMLX QVT.mdl plugin comprises just one source
> model folder of
> 9 files and three generated folders for emof, ecore models. I
> propose that these plugins be hosted by QVT Declarative as
>
> org.eclipse.qvt
> org.eclipse.qvt.importer.rose
>
> Since the intent of org.eclipse.qvt is to provide a useful
> bug-fix of the OMG distribution, it seems sensible to provide
> the full bug-fix in one QVT Declarative location rather than
> splitting it across QVT Declarative and QVTOML.
>
> Comments?
>
> Regards
>
> Ed Willink
>
>
> _______________________________________________
> m2m-dev mailing list
> m2m-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/m2m-dev
>