Community
Participate
Working Groups
As reported in the EMF newgroup "EAnnotations and their contents" thread, the name of for instance a Collection can be computed prematurely by ill-advised user code, causing a problem when the name is set later in the load process. As observed by Ed Merks, the cached value does not track its algorithmic source. ---- Specification-wise, OclExpression::name is a mess, caused by the dubious inheritance of OclExpression from TypedElement. What is the name of "1 + 2"? unspecified. The name "Set(XX)" is inadequate because there may be multiple XXs (a known OCL issue) but specigfied as derived. The name in x->forAll( is of course a normal name. ---- MDT/OCL is stuck with the OMG specification, but should at least cache the name in a way that does not violate EMF behaviours.