Re: [abc] Verification error

Are you suggesting that we use jlo-rt as a benchmark in the oopsla paper?
It could be interesting, but perhaps a bit confusing for the reader to
have one of the benchmarked systems as a benchmark itself.

-O

On Mon, 23 Jan 2006, Eric Bodden wrote:

> Hi.
>
> So far I got now almost everything working. The only remaining problem
> is the following, and I think it's worth to be reported, because it
> seems to be a bug in Soot. When weaving Oege's Nullpointercheck aspect
> into jlo-rt.jar, I get the following error:
>
> Exception in thread "main" java.lang.VerifyError: (class:
> rwth/i2/ltlrv/data/Weak
> ValuesHashMap$1, method: <init> signature:
> (Lrwth/i2/ltlrv/data/WeakValuesHashMap
> ;Ljava/util/Map$Entry;Ljava/lang/Object;)V) Expecting to find
> object/array on sta ck
> at
> rwth.i2.ltlrv.data.WeakValuesHashMap.entrySet(WeakValuesHashMap.java:1
> 77)
> at java.util.AbstractMap$1$1.<init>(AbstractMap.java:376)
> at java.util.AbstractMap$1.iterator(AbstractMap.java:375)
> at
> rwth.i2.ltlrv.afastate.impl.Proposition.specializeBindings(Proposition
> .java:410)
> at
> rwth.i2.ltlrv.management.EvalPropositionWeakMapTest.<clinit>(EvalPropo
> sitionWeakMapTest.java:24)
> at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native
> Method)
> at
> sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructo
> rAccessorImpl.java:39)
> at
> sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingCo
> nstructorAccessorImpl.java:27)
> at
> java.lang.reflect.Constructor.newInstance(Constructor.java:494)
> at junit.framework.TestSuite.createTest(TestSuite.java:131)
> at junit.framework.TestSuite.addTestMethod(TestSuite.java:114)
> at junit.framework.TestSuite.<init>(TestSuite.java:75)
> at junit.framework.TestSuite.addTestSuite(TestSuite.java:101)
> at
> rwth.i2.ltlrv.management.EvalEvaluationTests.suite(EvalEvaluationTests
> .java:27)
> at general.AllTests.suite(AllTests.java:19)
> at perf.Main.main(Main.java:86)
>
> Could that have to do with Junit using reflection? But even then, I
> think abc/Soot should not generate invalid bytecode.
>
> It's not urgent, but it might be something we should have fixed.
>
> Cheers,
> Eric
>
> --
> Eric Bodden
> Sable Research Group, McGill University
> Montreal, Canada
>
>
Received on Tue Jan 24 09:23:37 2006