Two PRs have shown up that have different message ids, this must never happen. https://github.com/jbossas/jboss-eap/pull/209/files#L1R2455 https://github.com/wildfly/wildfly/pull/4126/files#L8R2454 So all messages must now be reviewed to ensure they are globally unique.
The PR 4126 was never merged to Wildfly, it was superseded by PR4156 which put a few PRs together and changed the message ids as well. Thus, this particular case seems to be fine. We need, however, to check that all message ids in EAP are unique and in sync with Wildfly.
There is also a duplicity between JSFMessages.java of jsf.injections and jsf.subsystem, but these messages are the same, at least at the moment they are. Not sure whether it is a bug or not, see an example: jsf/injection/src/main/java/org/jboss/as/jsf/JSFMessages.java: @Message(id = 12655, value = "@ManagedBean is only allowed at class level %s") jsf/subsystem/src/main/java/org/jboss/as/jsf/JSFMessages.java: @Message(id = 12655, value = "@ManagedBean is only allowed at class level %s") And another duplicity between jdr and security subsystems which seems to be a real bug, see: https://issues.jboss.org/browse/WFLY-1626
There is a jira for JSF and other duplicated messages in Wildfly, see https://issues.jboss.org/browse/WFLY-1631 And 6.1.x's bugzilla for the JSF ones which are related to EAP: https://bugzilla.redhat.com/show_bug.cgi?id=980823
Closing this. See comment#8. There are no more collisions between EAP X WF messages and remaining issues are covered by 980823 and 901001.