I haven't studied the specs on this yet.
Sun's example clearly shows the entity bean calling ejbRemove() from
a business method in the entity bean itself. I would definitely
expect this to have undefined behavior since it is a lifecycle method
intended to be called from remove on one of the interfaces. However,
I haven't looked for a definitive statement in the spec yet, and I
suspect there isn't one.
I would expect that calling remove on the appropriate interface,
after marking the entity bean reentrant, should work. It's not clear
to me if you have tried this yet, could you clarify?
In any case crashing the server as a result of calling ejbRemove()
directly seems to me to be a bug. Could you file a jira for this
with as many details as possible?
many thanks,
david jencks
On Jul 12, 2006, at 5:35 AM, mahu2425 wrote:
Hello Dain,
I couldn't find it in the specs either, but I think it makes sense,
that an entity bean can remove itself.
An example, that an entity bean can remove itself is also given at
Sun's J2EE tutorial:
http://java.sun.com/j2ee/1.4/docs/tutorial/doc/JMSJ2EEex3.html
I my opinion this feature should be added in the next version of
Geronimo. I also found a few posts in the internet, that this
problem occured on several other J2EE server and has been fixed there.
Best regards,
Markus
Benachrichtigung bei E-Mail Empfang! - http://mail.lycos.de/app/
lycosinside/setupLI.exe
