Thanks Jason, that answers my question. I'll rework/rethink that PR to
account for the 3.x client behaviour.
-Jaikiran
On 12/09/17 11:22 AM, Jason Greene wrote:
Hi Jaikiran,
You probably already noticed alot of this, but just to make sure
everyone is on the same page the legacy ejb client is for the 3.x
branch of the ejb client code based, and the current client is the 4.x
branch. The 3.x branch is the same as the 2.x branch, but with updates
to synchronize dependencies and integrate with elytron capabilities.
So 3.x is intended to be 100% API compat with 2.x. 4.x on the other
hand has dropped some elements of the API that no longer make sense
and added new APIs. So 3.x is unlikely to be enhanced further, with
just bug fixes, and all new work going into 4.x. Since most access
patterns port without issue, we expect 3.x will have fairly low usage,
and will one day be phased out.
So basically the only reason to add test coverage with 3.x use
(legacy-ejb-client) is if there is a bug that affects it, or a change
was made that is likely to impact its operation. Since 4.x is protocol
compat with 3.x though, changes on the 4.x server side do need to
factor in 3.x client behavior.
-Jason
> On Sep 12, 2017, at 12:37 AM, Jaikiran Pai <jai.forums2013(a)gmail.com
> <mailto:jai.forums2013@gmail.com>> wrote:
>
> I was working on a change and a testcase related to one of the JIRAs
> open in WFLY and happened to notice that we now have a
> testsuite/integration/legacy-ejb-client module. It looks like this
> module has testcases that are also part of testsuite/integration/basic -
> i.e. copy pasted. So we now have 2 copies of a testcase for example
> EJBClientAPIUsageTestCase. What's the guideline when dealing with these
> testcases? Should both these testcases be kept up-to-date with any
> changes to the behaviour of the server interaction with the remote
> client? The context of my question is this PR
>
https://github.com/wildfly/wildfly/pull/10466
>
> -Jaikiran
>
> _______________________________________________
> wildfly-dev mailing list
> wildfly-dev(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/wildfly-dev
--
Jason T. Greene
Chief Architect, JBoss EAP
Red Hat