[keycloak-dev] Splitting subsystem into two

Stan Silvert ssilvert at redhat.com
Tue Jan 6 08:12:52 EST 2015


You may be right.  I think the decision comes down to answering the 
question, "Is it ever advantageous to have the server and adapter in the 
same classloader space (module)?"  The answer will hopefully become 
clear once I get a chance to talk to you in more detail about seamless 
adapter configuration.

I've got the client part working where a template configuration is 
applied to any unsecured WAR.  We just need to decide how the subsystem 
will register the application with the server.  That's the part where it 
might be very helpful to have both in the same module so we can just 
make direct calls.  Even so, we can't really do that today without 
significant refactoring.

On 1/6/2015 7:48 AM, Stian Thorgersen wrote:
> https://issues.jboss.org/browse/KEYCLOAK-872 made me think it might be cleaner to split the subsystem in two, one for the server and another for the adapter.
> _______________________________________________
> keycloak-dev mailing list
> keycloak-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/keycloak-dev


More information about the keycloak-dev mailing list