I have created TUSCANY-3683 and submitted a patch for DefaultRMIHost.java.
-- Sebastian

> -----Original Message-----
> From: Simon Nash [mailto:[email protected]]
> Sent: Wednesday, September 22, 2010 12:21 AM
> To: [email protected]
> Subject: Re: FW: FW: RMI cannot be promoted
> 
> Millies, Sebastian wrote:
> > the error occurs when the composite with the promoted binding
> > is activated in isolation, independent of whether it is used
> > as a component implementation.
> >
> > The service name that is registered twice in the rmiregistry
> > is the serviceName attribute of the binding.rmi element, in
> > this case "LogAdapterServiceRMI".
> >
> Using the service name as the rmiregistry name isn't a very good idea.
> It won't work if two different components happen to have RMI services
> with the same name.
> 
> > One way to work around the error would be to provide an explicit
> > RMI binding for the promoted service using a new RMI-name, e. g.:
> >
> >  <service name="SLF4JLogAdapterService"
> promote="SLF4JAdapterComponent/LogAdapterService">
> >     <ts:binding.rmi host="localhost" port="1099"
> serviceName="PromotedLogAdapterServiceRMI"/>
> >   </service>
> >
> > But this seems somewhat counterintuitive - if no binding is specified
> for a
> > promoted service, the binding for the component service should simply
> > be taken from the original service definition. So I would still
> classify Tuscany's
> > behaviour as a minor bug.
> >
> I agree.  Would you like to raise a JIRA and perhaps submit a fix?
> 
>    Simon
> 
> > -- Sebastian
> >
> >> -----Original Message-----
> >> From: Simon Laws [mailto:[email protected]]
> >> Sent: Tuesday, September 21, 2010 11:24 AM
> >> To: [email protected]
> >> Subject: Re: FW: RMI cannot be promoted
> >>
> >> On Mon, Sep 20, 2010 at 3:16 PM, Millies, Sebastian
> >> <[email protected]> wrote:
> >>> by the way, the obvious thing (it seems to me) would be to use
> >>> Naming.rebind() instead of Naming.bind(). Is there a particular
> >>> reason why Naming.bind() is used? -- Sebastian
> >>>
> >>>
> >>>> -----Original Message-----
> >>>> From: Millies, Sebastian [mailto:[email protected]]
> >>>> Sent: Monday, September 20, 2010 1:10 PM
> >>>> To: [email protected]
> >>>> Subject: Bug: RMI cannot be promoted
> >>>>
> >>>> Hello there,
> >>>>
> >>>> is this a known bug in Tuscany 1.6? Couldn't find it in JIRA,
> >>>> but maybe my search was not good.
> >>>>
> >>>> A service with an RMI binding cannot be promoted, because when
> >>>> the binding for the promoted service is processed, the Tuscany
> >>>> CompositeActivatorImpl attempts to start the
> >> RMIServiceBindingProvider
> >>>> a second time, which leads to the DefaultRMIHost wanting to bind
> the
> >>>> service under the same name in the RMI registry, which produces an
> >>>> AlreadyBoundException.
> >>>>
> >>>> Here's a snippet from the composite.xml:
> >>>>
> >>>> <?xml version="1.0" encoding="UTF-8" standalone="no"?>
> >>>> <composite xmlns="http://www.osoa.org/xmlns/sca/1.0";
> >>>> xmlns:ts="http://tuscany.apache.org/xmlns/sca/1.0";
> >>>> name="logger_composite"
> targetNamespace="http://ps.softwareag.com/";>
> >>>>   <component name="SLF4JAdapterComponent">
> >>>>     <implementation.java
> >>>>
> >>
> class="com.softwareag.ps.platform.logadapter.impl.SLF4JAdapterImpl"/>
> >>>>     <service name="LogAdapterService">
> >>>>       <interface.java
> >>>>
> >>
> interface="com.softwareag.ps.platform.logadapter.LogAdapterService"/>
> >>>>       <binding.sca/>
> >>>>       <ts:binding.rmi host="localhost" port="1099"
> >>>> serviceName="LogAdapterServiceRMI"/>  <!-- *** -->
> >>>>     </service>
> >>>>   </component>
> >>>>   <service name="SLF4JLogAdapterService"
> >>>> promote="SLF4JAdapterComponent/LogAdapterService"/> <!-- *** -->
> >>>> </composite>
> >>>>
> >>>> -- Sebastian
> >>>>
> >> Hi Sebastien
> >>
> >> Is the promoted service being used by a component implemented using
> >> implementation.composite? What is the service name that it thinks is
> >> being registered twice?
> >>
> >> Simon
> >>
> >> --
> >> Apache Tuscany committer: tuscany.apache.org
> >> Co-author of a book about Tuscany and SCA: tuscanyinaction.com

Reply via email to