Community
Participate
Working Groups
I have spent quite some time on many occasions trying to fathom "Cannot find compatible feature" ... "in sealed EClass", and after blaming the tool, I have eventually found a meta-model issue. Each of the following could be distinctly diagnosed: feature has incompatible containment feature has incompatible multiplicity ??feature has incompatible transient ??feature has incompatible volatile ??feature has incompatible abstract ??feature has incompatible interface With the above clues, I could have solved my problems much more rapidly. [Further to Bug 286053, it might be also be kind to run a validation on the source meta-models as part of the diagnosis of an incompatible feature.]
Another nice obscure problem: Ecore model includes e.g. platform:/plugin/org.eclipse.xtext.common.types/model/JavaVMTypes.ecore but Xtext imports platform:/resource/org.eclipse.xtext.common.types/model/JavaVMTypes.ecore or http://www.eclipse.org/xtext/common/JavaVMTypes
And another: A1 extends B, C A2 extends B, C then aB=(A1,A2) gives the same message if aB requires a B, but does not allow a C, seemingly because the last superType C, is checked fpr comaptibility with B rather than any.
(In reply to comment #1) See Bug 323674 for a similar problem between platform:/resource/... and ../../... access.
And another obscure one: If grammar B extends grammar A and both grammar A and grammar B have a reference to feature N of type X. Then the meta-model is changed so the feature N is of type Y. Editing grammar B shows the error but once corrected it is still shown as an error until the same error in the base grammar is also corrected. In this scenario surely the base grammar should have a red squiggle similar to JDT's 'type hierarchy is inconsistent' diagnosis.
Scheduled for M5 as grammar mixin and Ecore import will become more popular with Xbase.
Improved error messages.
Closing all bugs that were set to RESOLVED before Neon.0
