Thank you Jochen and Paul for looking into it. At first I’ve tried this exact code and it fixes my test case, but then it differs from the code in AsmReferenceResolver only by logic for inner classes, and I have strong feeling that the exact same code should be reused. I’m leaving this up to you.
In the meanwhile I’ve added an ability to disable ‘asmResolving’ via optimization options map in IntelliJ. — Daniil Ovchinnikov JetBrains > On 15 Jan 2019, at 11:38, Paul King <pa...@asert.com.au> wrote: > > The tests still pass replacing the try/catch with: > ClassNodeResolver.LookupResult lookupResult = > getClassNodeResolver().resolveName(name, CompilationUnit.this); > return lookupResult == null ? null : lookupResult.getClassNode(); > Not sure whether returning null will cause a problem but if so it isn't > triggered by our current test suite. > > Cheers, Paul. > >
