Community
Participate
Working Groups
This planning bug aggreagtes issues that require a Pivot metamodel change, and so probably cannot be done without a major version change,
Bug 509308 residue: ShadowExp::value is dead ShadowExpCS::value is dead
Bug 509713#c2, review InvalidEvaluationException, InvalidValueException, RuntimeException inheritance.
Bug 514353 UML's Class::nestedClassifiers needs NestedClass
Why is Element::ownedAnnotations of type Element not Annotation?
Bug 341944 Add transient FeatureCallExp.implementation
Use UML's Expression instead of LanguageExpression
Fold TemplateParameterSubstitutionHelper into LibraryOperation
See also Bug 519896. Rationalize Iteration/Operation. Operation::bodyExpression => Operation::ownedBody Iteration::ownedIterators => Operation::ownedParameters + ?? isIterator/IteratorParameter Iteration::ownedAccumulators => Operation::ownedParameters + ?? isAccumulator/AccumulatorParameter
Bug 528355 identifies that PivotObjectImpl.esObject should be an Object to support Ecore pre/post-cpnditions.
Bug 530375 identifies that distinct Property::explicitOpposite and Property::implicitOpposite are needed to avoid a custom serialization of the logically transient implicit opposite.
Residue of Bug 532613. Eliminate MorePivotable, Nameable, Pivotable, Visitable, VisitableCS from the ecore/uml models.
(In reply to Ed Willink from comment #8) > Operation::bodyExpression => Operation::ownedBody No. See Bug 481664 for QVTb Function considerations.
Eliminate the deprecated Precedence.order/associativity and ExpCS.precedence.
Bug 543187 identifies that TemplateParameter.constrainingClasses is a horrible rag-bag. Redesign.
Distinct CollectionType.elementType / MapType.keyType/valueType are a pain. Migrate to IterableType.elementType + MapType.valueType.
Bug 546497 replaces OppositePropertyCallExp by PropertyCallExp.isOpposite
Bug 551822 suggests making ExpressionInOCL less redundant when used by Operation.
Bug 561265 identifies that NullValue should extend MapValue, OclVoidTypeId should extend MapTypeId.
Bug 520440 identifies that the selectively boxed return values from the evaluator make a richer SymbolicBooleanValue impossible. Let everything be optionally boxed.
Bug 561941 kicks off OCL 3 where many of these chnages may happen.
Bug 520440 is a run-time not metamodel issue.
Bug 566722 identifies that TemplateParameter.constrainingClasses should be constrainingTypes. From the Bug 499371 duplicate: Bug 492801: Move ExpressionInOCL from LanguageExpression derivation to CompleteClass cache element. CollectionType.elementType is currently changeable, derived. Change to readonly, derived.
*** Bug 493371 has been marked as a duplicate of this bug. ***
Bug 565412 considers how a more Xtext-friendly AS/CS might support persistence of arbitrary comments.
Add known owners such as for CollectionLiteralPart, MapLiteralPart, ShadowPart, TupleLiteralPart
Make TypeExp a LiteralExp
Bug 574430 provides an approach to eliminate CompleteClass etc. Bug 574431 provides an approach to simplify behavioral types.
Add missing CoIteratorVariable
Bug 574884 introduces a Union type. ?? rather similarly an Alias type for Ord(T) as OrderedSet(T).
Bug 575157 identifoes that LambdaType should be an aggregate of Parameter
Add IteratorVariable.owningLoop etc. (Use Let/Iterator/Parameter Variable consistently.)
Bug 577242 advocates adding a Map EntryType and making MapLiteralPart a TypedElement.
Bug 482583 advocates changing allInstances to an ordered return.
Bug 578010 suggests that UML stereotype support such as Class::extenders should be normaliuzed away so that only regular Property are needed.
Bug 578994 motivates a PackageOwner abstraction of Model and Package.
Bug 579109 identifies the Element.ownedExtensions is yet another property that shoud be transient.
(In reply to Ed Willink from comment #31) > Add IteratorVariable.owningLoop etc. (Use Let/Iterator/Parameter Variable > consistently.) Bug 580356 identifies a confusion when "this" was added for QVTd; the ParamterVariable was omitted. There can be a very strong Paramter/Variable distinction eliminating the VariableDeclaration pragmatisms.
Bug 580407 identifies that the bloat of TemplateParameterSubstitution / TemplateBinding could be much simplified. Probably by positionally equivalent TemplateSignature.ownedParameters: Sequence(TemplateParameter) TemplateSignature.bindings : Sequence(Class) rather like EGenericType.
Bug 580511 identifies that the binding of a TemplateParameter should be split off to TemplateParameterType leaving TemplateParameter as a TypedElement declaration.
Bug 580662 cleans up implicit/unnavigable opposite terminology
Bug 580758 resolves the bloat associated with CollectionItem / CollectionLiteralPart and accommodates the pathlogical meta-case of a collection of CollectionRange.
Bug 580755 migrates the Pivot mastership from the *.uml files in the build plugin to PivotStructural.ecore in the pivot plugin.
(In reply to Ed Willink from comment #17) > Bug 551822 suggests making ExpressionInOCL less redundant when used by > Operation. Bug 580356 goes further advocating eliminating the unhelpful duplocation of ParamterVariablle and representedParamter.
Bug 581882 identifies that the *Value interfaces are not assignable to java.lang.Number without a case. Fold the *ValueImpl into *Value.
Bug 582029 introduces a self=true EAnnotation for a Property to workaround the no-cycle eOpposite Ecore limitation.
BUg 582444 argues that a Constraint should be a derived Operation.