Community
Participate
Working Groups
In the /org.eclipse.qvtd.build.etl.tests/src/org/eclipse/qvtd/build/etl/tests/UmlToRdbms/UmlToRdbms.qvtc example, for the primitive type MR mappings (e.g. booleanToBoolean_MR), there is a dependency to PrimitiveToName.owner. Since the mapping has a variable of a specific subtype, e.g. BooleanToBoolean, the dependency should be shown as BooleanToBoolean.owner .
Similarly, in the opposite case, i.e. the mapping has an input of a type which has subtypes, the property dependencies should be to the given property of all subtypes. Yes, the types are connected by inheritance edges, but this is not true for the properties. This causes propertyes form (abstract) supertypes to be shown in the graph only as requirements, but there is no mapping that produces them.
(In reply to Horacio Hoyos from comment #1) > Similarly, in the opposite case, i.e. the mapping has an input of a type > which has subtypes, the property dependencies should be to the given > property of all subtypes. Yes, the types are connected by inheritance edges, > but this is not true for the properties. This causes propertyes form > (abstract) supertypes to be shown in the graph only as requirements, but > there is no mapping that produces them. Disregard this comment please. The behaviour is the same as originally posted, but this is the effect it has on productions.
