Community
Participate
Working Groups
Maybe the following stack is sufficient in order to reproduce? java.lang.NullPointerException at org.eclipse.objectteams.otdt.internal.ui.text.correction.MappingProposalSubProcessor.createMethodSpec(MappingProposalSubProcessor.java:404) at org.eclipse.objectteams.otdt.internal.ui.text.correction.MappingProposalSubProcessor.createCalloutFromInferred(MappingProposalSubProcessor.java:395) at org.eclipse.objectteams.otdt.internal.ui.text.correction.MappingProposalSubProcessor.getMaterializeInferredCalloutsInheritedProposal(MappingProposalSubProcessor.java:322) at org.eclipse.objectteams.otdt.internal.ui.text.correction.QuickFixProcessor.process(QuickFixProcessor.java:396) ...
Put on my radar.
Although I can somewhat narrow down the issue I fail to find a way to reproduce. What we know: - we have a callout binding inferred from an inherited abstract method -> this implies that the inherited role method and a corresponding base method exist and match in name and all parameter types (I tried with missing types, but then inference will abort) - we are computing a proposal to materialize that callout - in l. 395 "mapping.getRoleMethod()" evaluates to null Most probable track: - while parameters were OK, perhaps the return type of role and/or base method was unresolved / missing. - in this case DefaultBindingResolver.getMethodBinding() could possible return null, *but only if* (!this.isRecoveringBindings). - this again is difficult to achieve, perhaps by a JME during CompilationUnitResolver.resolve (would indicate an inaccessible JavaProject) *or* an ICE in that method. Apparently too many unexpected things need to meet to make this bug happen, so analysis without a witness gives no definite result. Closing as WORKSFORME.