[jboss-jira] [JBoss JIRA] Closed: (JBREM-61) Possible RMI exception semantic regression

Wed Sep 26 05:31:02 EDT 2007

     [ http://jira.jboss.com/jira/browse/JBREM-61?page=all ]

David Lloyd closed JBREM-61.
----------------------------


> Possible RMI exception semantic regression
> ------------------------------------------
>
>                 Key: JBREM-61
>                 URL: http://jira.jboss.com/jira/browse/JBREM-61
>             Project: JBoss Remoting
>          Issue Type: Bug
>          Components: transport
>            Reporter: Scott M Stark
>         Assigned To: Tom  Elrod
>             Fix For: 1.0.1 final
>
>
> The org.jboss.test.security.test.EJBSpecUnitTestCase.testDeepRunAs method is currently failing in head after the switch of the default invoker to the remoting based unified invoker due to a subtle difference in how exceptions are handled by the jrmp invoker. The test is expecting that the java.rmi.AccessException thrown by the ejb contaner be wrapped in a java.rmi.ServerException. This is a wrapping exception used by the jrmp implementation. The java.rmi.ServerException javadoc:
> A ServerException is thrown as a result of a remote method invocation when a RemoteException is thrown while processing the invocation on the server, either while unmarshalling the arguments, executing the remote method itself, or marshalling the return value. A ServerException instance contains the original RemoteException that occurred as its cause.
> The unified invoker is simply throwing back the java.rmi.AccessException. Although this looks like what the JRMPInvoker is doing, this ignores the fact that the jrmp implementation sits between the network and the client. We probably need a strictRMIException mode that allows for the same behavior.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira