That was the trick with link(), if you have the child in classloader C and the parent in C.parent then you get the parent included. This is why i think something can have been broken.
Romain Manni-Bucau @rmannibucau <https://twitter.com/rmannibucau> | Blog <https://rmannibucau.metawerx.net/> | Old Blog <http://rmannibucau.wordpress.com> | Github <https://github.com/rmannibucau> | LinkedIn <https://www.linkedin.com/in/rmannibucau> | Book <https://www.packtpub.com/application-development/java-ee-8-high-performance> Le mer. 16 mai 2018 à 07:43, Mark Struberg <strub...@yahoo.de.invalid> a écrit : > Yes, but the weird thing is that it comes from the parent CL, so the > scanner has not too much to do with it. > Btw, by excluding the Earname test I came a bit further. > > Now the following are failing in the RS module > [ERROR] Errors: > [ERROR] ApplicationStarTest.checkStarIsNotAnIssue:43 » > InternalServerError HTTP 500 In... > [ERROR] DisableTomEEJohnzonTest.client:71 » InternalServerError HTTP 500 > Internal Serv... > [ERROR] DisableTomEEJohnzonTest.server:66 » IO Server returned HTTP > response code: 500... > [ERROR] JSonStreamingOutputTest.run:51 » InternalServerError HTTP 500 > Internal Server ... > [ERROR] ProviderWithoutAnnotationTest.run:62 » InternalServerError HTTP > 500 Internal S... > [ERROR] TomEEConfigurableJohnzonTest.run:91 » IO Server returned HTTP > response code: 5... > [INFO] > > > > But this is likely to a glitch in new bval.Or due to me being de_AT > located... ;) > LieGrue,strub > > On Wednesday, 16 May 2018, 06:51:59 CEST, Romain Manni-Bucau < > rmannibu...@gmail.com> wrote: > > Mark > > Previous test issue can be due to xbean upgrades or scanning changes, did > you check that? We tend to link implicitly the finder for these tests but > can have been broken at some point. > > Le mar. 15 mai 2018 23:16, Mark Struberg <strub...@yahoo.de.invalid> a > écrit : > > > fb_tomee8 > > I've fixed this now by fixing the broken bean. Wonder how this could have > > ever worked.Might also be partly random due to different build order? > > > > But now I'm off to the next bug. > > EarModuleNamesTest > > This one also scans for a resource due to haing the openejb-itests-beans > > on the classpath.But since the config is done manually it misses the > TxMgr > > setup, which is null. > > I exclude this test for now and want to see how far I come with the other > > tests. > > LieGrue,strub > > > > On Tuesday, 15 May 2018, 22:03:03 CEST, Jonathan Gallimore < > > jonathan.gallim...@gmail.com> wrote: > > > > Which branch? > > > > Jon > > > > On Tue, May 15, 2018 at 9:51 PM, Mark Struberg <strub...@yahoo.de.invalid > > > > wrote: > > > > > Well, this looks weird anyhow > > > @Stateless > > > public class AnnotatedFieldInjectionStatefulBean { > > > > > > compare the name and the annotation.... > > > And the additional error messages > > are:org.apache.openejb.config.ValidationWarning: > > > @Init is ignored for beans of type Stateless. Class: > > > org.apache.openejb.test.stateful.AnnotatedFieldInjectionStatefulBean > > > Method: create" > > > "org.apache.openejb.config.ValidationWarning: @Remove is ignored for > > > beans of type Stateless. Class: org.apache.openejb.test.stateful. > > > AnnotatedFieldInjectionStatefulBean Method: remove"" > > > > > > Well, yes, true! > > > Is this a negative test bean? ^^ > > > > > > On Tuesday, 15 May 2018, 21:48:26 CEST, Mark Struberg > > > <strub...@yahoo.de.INVALID> wrote: > > > > > > hi folks! > > > I get 2 unit test failures which look completely weird to me.Plus they > > > pass in the IDE but fail in mvn. > > > ERROR] Failures: > > > [ERROR] CheckDescriptorLocationTest.testWarWithDescriptorInMetaInf > Keys > > > do not match expected:<...or.incorrectLocation[]> but > > > was:<...or.incorrectLocation[ > > > ignoredMethodAnnotation > > > ignoredMethodAnnotation]> > > > [ERROR] CheckDescriptorLocationTest.testWarWithDescriptorInRoot Keys > do > > > not match expected:<...or.incorrectLocation[]> but > > > was:<...or.incorrectLocation[ > > > ignoredMethodAnnotation > > > ignoredMethodAnnotation]> > > > > > > > > > The unit test in question is > > > org.apache.openejb.config.CheckDescriptorLocationTest# > > > testWarWithDescriptorInRoot > > > This bundles a single FooBean.class with a single error.We get this > > error, > > > but also a few other test failures which apparently should not happen. > > > Those come from beans in transitive dependency projects like > > > openejb-itests-beans. > > > It passes in my IDE since there the classpath is different... > > > Any idea what happened? > > > LieGrue,strub > > > > >