SessionContext.getInvokedBusinessInterface Regression

[alrubinger]

http://jboss.hudson.alrubinger.com/job/EJB3_Integration_TestSuite_AS_TRUNK/lastBuild/testReport/org.jboss.ejb3.test.ejbcontext.unit/EjbContextUnitTestCase/testStatelessInvokedBusinessInterface/

https://jira.jboss.org/jira/browse/EJBTHREE-2076

I've got this working locally with some changes to core and Carlo's ejb3-context component on GitHub.  In fact I finally went into my own account and forked his repo to make a pull request for:

http://github.com/ALRubinger/jboss-ejb3-context/commit/98fa6f0f7966ea8b83d36866d7238c3b52fe9de5

This commit is a little wonky because I accidentally added a bunch of metadata files in there, so when you guys get online someone's gonna have to help me figure out how to revert a change on my remote copy and local copy as I learn DVCS a bit.

S,

ALR

[wolfc]

The invocation InvocationContext is already authoritative on the answer of getInvokedBusinessInterface(). So the extra check is not needed at all. Although we could argue about the message in the exception itself.

http://mike.lab.bos.redhat.com:8380/hudson/job/JBoss-AS-6.x-tests-ejb3/481/testReport/org.jboss.ejb3.test.ejbcontext.unit/EjbContextUnitTestCase/

[wolfc]

Reverting a remote repo is a bit tricky in the sense of it already being published.

I did 'merge' your commit into my repo using the 'Fork Queue', but it doesn't really give the desired result. I would rather see commits merged in instead of recommitted. I did a merge into branch alr2 just for comparison.

Now that I have a reference to your tag we can see what happens if we don't play nice.

Now it's a 'pick your poison' game.

If you don't care about keeping a reference to that commit, just do

$ git reset HEAD^

do fiddling and commit again.

The push will result in a failure, so you'll need to remove the remote reference.

See http://github.com/guides/remove-a-remote-branch