Zhi Liu wrote:
> Hi Ken,
>
> thank you very much for your prompt wonderful explanation.
> If you have time ,maybe you can answer one more question from me:
> Is that through the remote interface the Dynamic RMI-IIOP creates
> thenecessary RMI stubs ?
In the case of an EJB 3.0 Remote business interface, yes. For 2.x (and
earlier) access, it's the Home and Remote component interfaces.
--ken
>
> thank you again
>
> ---Zhi
>
> ------------------------------------------------------------------------
> *From:* Kenneth Saks [mailto:Kenneth.Saks_at_Sun.COM]
> *To:* Zhi Liu [mailto:zhi.liu_at_thoranet.de]
> *Cc:* ejb_at_glassfish.dev.java.net
> *Sent:* Fri, 14 Jul 2006 15:59:28 +0200
> *Subject:* Re: must remote interface class in classpath of the
> EJB3 remote client ????
>
> Zhi Liu wrote:
>
>>
>> ---------------------------------------------------------------------------------------------
>> If I include the class of remote interface in classpath, It
>> works. But in your EJB FAQ you didn't mention that. It said in
>> EJB FAQ :
>> /Our implementation uses a feature called "Dynamic RMI-IIOP" that
>> creates any necessary RMI stubs at runtime in a way that is
>> completely hidden from the application./
>
>> If I don't misunderstand with the Dynamic RMI-IIOP Feature I
>> don't need the stub and don't need the remote interface class in
>> classpath, is that right?
>
> Hi Zhi,
>
> You do need to add the Remote interface to the stand-alone java
> client.. Dynamic RMI-IIOP solves a different problem, which is
> that in many implementations there are stubs based on that
> interface that must be statically generated and also added to the
> classpath. We'll update the wording a bit to clarify this point.
> Thanks for your feedback.
>
> --ken
>
>>
>> Thank you for you helf
>>
>> best regards
>>
>> zhi
>
>
>
>
>