I have now seen that the archieve scanner used by hibernate in in widlfly comes from:
Could it be that wildfly has implemented such a scanning extension for eclipselink?
And I am now looking at the fact that wildlfy also offers a
I have not looked at the source of this eclipsleink archieve, but My guess is that widlfly has implemented proper support for eclipselink, and there is only some module configuration that I must have incorrectly set when using eclipselink.
Any pointers would be appreciated.
Just to report back - I was correct in my assumption that Jboss hs already created an eclipselink enhancement to deal with the virtual file system.
The problem was that the wildfly configutation I was working on was probably created based on old wildfly documentation and it did not include as a resource the:
Manve library as above.
In addition the system property to tweak the eclipseli link archieve factory was not configured in the domain, so the default archieve factory of eclipselink was being used, which was incorrect.
Sorry for the question ... this was a simple matter of someone configuring things based on stale documentation.
One last point the relevant references for this are:
2. Module configuration as follows:
<module xmlns="urn:jboss:module:1.1" name="org.eclipse.persistence">
<exclude path="javax/**" />
3. System property:
<property name="eclipselink.archive.factory" value="org.jipijapa.eclipselink.JBossArchiveFactoryImpl"/>
in the stand-alone,xml
Works like a charm.