well, here a patch which solve it: https://gist.github.com/35bf4109848562fd1221
it let the unsetThreadLocal but rebind the injection point when we'll do a getreference with an empty stack of injection point. - Romain 2012/2/22 Romain Manni-Bucau <[email protected]> > why in org.apache.webbeans.inject.AbstractInjectable#inject the injection > point is removed at the end? > > (through InjectionPointBean.unsetThreadLocal();) > > it seems it is the cause of this regression > > - Romain > > > > 2012/2/20 Rohit Kelapure <[email protected]> > >> This could be due to recent changes made in >> https://issues.apache.org/jira/browse/OWB-645 >> I will investigate and get back. >> >> -Rohit >> >> >> On Fri, Feb 17, 2012 at 8:30 AM, Romain Manni-Bucau >> <[email protected]> wrote: >> > Hi, >> > >> > it seems recent change(s) introduced a regression on TCKs: >> > >> > Failed tests: >> > >> testGetInjectableReferenceOnBeanManager(org.jboss.jsr299.tck.tests.lookup.injectionpoint.InjectableReferenceTest) >> > >> > any idea? >> > >> > - Romain >> > >
