The test reproduces the abstraction and the registration does not change the code path so it can be a "bad stack" which would make the OWB BeanMetadataProducer not finding the bean?
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 lun. 17 déc. 2018 à 18:09, Laird Nelson <[email protected]> a écrit : > On Mon, Dec 17, 2018 at 12:12 AM Romain Manni-Bucau <[email protected]> > wrote: > >> Hello Laird, >> >> I didn't manage to reproduce with this test: >> > > Yes, but that test is different from the structure and usage of the > Narayana transaction componentry in many ways. For example, the Narayana > extension adds the interceptor types as annotated types (i.e. does not call > addInterceptor()), and the interceptedBean field is in a base class, not > the actual interceptor. I don't know why (or if) this would make a > difference, of course, but using the Narayana extension does indeed fail > under OpenWebBeans. I'll continue digging. > > Best, > Laird >
