I think you hit the nail on the head with this one. As I mentioned, I was
throwing java.lang.IllegalArgumentException in one business method in both
of the classes in question. I commented out throwing that exception and
all jboss verifier messages went away. So are you saying this is a bug in
jboss verifier? I don't like the idea of not declaring the exception. I
went ahead and replaced the IllegalArgumentException with a new one
extending org.omg.CORBA.UserException. Are you saying that the exception
class could just extend java.lang.Exception and that would be
fine? Anyway, thanks for the help. I had already rooted through the
verifier code and had come up with a way to make the error messages more
specific. I've never tried to modify and open source project and
participate as a developer. But I signed up with sourceforge and was about
to checkout a CVS sandbox of the JBoss 2.4 code. I'd like to learn more
about writing an application with JBoss, before trying to enhance it
(intended irony). Again, thank you everyone.
Fred.
At 02:35 PM 7/12/2001, you wrote:
>The RuntimeException constraint arises from an overzealous intrepretation
>of the Java-CORBA mapping spec by the jboss Verifier. The jboss Verifier
>considers a RuntimeException declared in a business method invalid because
>it has no CORBA mapping. True enough, but that's because none is needed.
>The mapping spec (OMG document ptc/99-03-09 Sec. 28.3.4.4) states:
>
>"unchecked exception classes are assumed to be mapped to the implicit
>CORBA system exception, and are therefore not explicitly declared in OMG IDL."
>
>The EJB spec allows throwing a RuntimeException from a business method,
>and in fact encourages it for certain well-defined situations (cf. Sec.
>17.2.2, ejb 2.0 spec). The only note on _declaring_ a RuntimeException is
>the following:
>
>"Note that the javax.ejb.EJBException is a subclass of the
>java.lang.RuntimeException, and therefore it does not have to be listed in
>the throws clauses of the business methods."
>
>There is no firm edict _not_ to declare a RuntimeException in the method.
>
>A RuntimeException declaration should be allowed by jboss. The work-around
>is simple--don't declare it--but the Verifier message does not lead one to
>this conclusion. I will file a jboss bug report on this.
>
>--
>Fred Loney
>Spirited Software, Inc.
>[EMAIL PROTECTED]
>
>Burkhard Vogel wrote:
>
> > Hi,
> > I thought in the same pattern, but its in the spec that RuntimeExceptions
> > are not valid.
> > Burkhard
> > ----- Original Message -----
> > From: "Hermann RANGAMANA" <[EMAIL PROTECTED]>
> > To: <[EMAIL PROTECTED]>
> > Sent: Thursday, July 12, 2001 5:58 PM
> > Subject: Re: [JBoss-user] More on RMI/IIOP warnings
> >
> > > I've had the same problem few days ago...
> > >
> > > Remove all EJBException from the throws clause of ALL of your business
> > > methods in the remote interface. If there's only ONE business methods
> wich
> > > declare EJBException in its "throws" clause, then that remote interface
> > > won't be a valid type of RMI/IIOP...
> > >
> > > Check out methods in your home interface, too ... and make sure they
> > aren't
> > > declared as throwing EJBException ...
> > >
> > > I don't really know why jBoss (or the EJB spec ...) disallows
> EJBException
> > > as part of exception list of a business method. Since EJBException
> extends
> > > RuntimeException extends Exception implements Serializable, i don't see
> > any
> > > reason why a class containing a method declaring EJBException should not
> > be
> > > a valid type of RMI/IIOP....
> > >
> > > May be someone can shade a light on this ..
> > >
> > > --hermann
> > >
> > > ----- Original Message -----
> > > From: "Maraya Michael" <[EMAIL PROTECTED]>
> > > To: <[EMAIL PROTECTED]>
> > > Sent: Thursday, July 12, 2001 4:44 PM
> > > Subject: RE: [JBoss-user] More on RMI/IIOP warnings
> > >
> > >
> > > >
> > > > AFAIK, The remote and home interfaces shouldn't be serializable,
> > > > just their parameters and arguments.
> > > >
> > > > I'm still trying to get the hang of this, too. Send me your home
> > > > and remote interfaces, your bean implementation and any referenced
> > classes
> > > > and let me see if I can help. Send them to [EMAIL PROTECTED]
> > > >
> > > > ---
> > > > Michael R. Maraya
> > > >
> > > > > ----------
> > > > > From: Frederick N. Brier[SMTP:[EMAIL PROTECTED]]
> > > > > Reply To: [EMAIL PROTECTED]
> > > > > Sent: Wednesday, July 11, 2001 11:51 PM
> > > > > To: [EMAIL PROTECTED]
> > > > > Subject: [JBoss-user] More on RMI/IIOP warnings
> > > > >
> > > > > SomeClass and AnotherClass are interfaces that extend EJBObject which
> > in
> > > > > turn extend java.rmi.Remote. I don't believe you are supposed to
> > define
> > > > > and interface as both Serializable and Remote.
> > > > >
> > > > > So these are Remote interfaces for classes which I assume are
> > > dynamically
> > > > > generated by JBoss and act as front end redirectors (for the Bean
> > > classes)
> > > > >
> > > > > whose stubs are the actual objects serialized and should be able
> to be
> > > > > passed and returned as parameters. These redirectors are associated
> > > with
> > > > > the EJB container and then feed into the chain of interceptors that
> > > > > execute
> > > > > prior to invoking the method in the bean. Maybe I'm babbling, but I
> > > think
> > > > >
> > > > > I'm getting the hang of the architecture. I'm just confused as
> to why
> > > the
> > > > >
> > > > > verifier is giving me this error.
> > > > >
> > > > > Is there a debug switch, log trace, or a suggested approach to trace
> > > into
> > > > > the JBoss verifier that could perhaps give me more insight into
> what I
> > > > > might be doing wrong? BTW, I did try Fred Loney's suggestion and
> > > removed
> > > > > all data member references of the EJB's EJBObject interface type. No
> > > > > change in the error messages. Please, HELP! Suggestions?!??
>
>
>_______________________________________________
>JBoss-user mailing list
>[EMAIL PROTECTED]
>http://lists.sourceforge.net/lists/listinfo/jboss-user
_______________________________________________
JBoss-user mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-user
