river-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Firmstone <j...@zeus.net.au>
Subject Re: Dirmi might be something to look at
Date Sun, 27 Jan 2013 01:29:35 GMT
On 27/01/2013 7:57 AM, Tim Blackman wrote:
> On Jan 26, 2013, at 8:26 AM, Peter Firmstone<jini@zeus.net.au>  wrote:
>> There are parts of Jini that depend on proprietary sun jvm namespace, preventing
it from compiling on other jvm's:
>> Compiling 863 source files to C:\Users\peter\Documents\NetBeansProjects\peterConcurrentPolicy\build\classes
>> C:\Users\peter\Documents\NetBeansProjects\peterConcurrentPolicy\src\com\sun\jini\phoenix\SunJrmpExporter.java:31:
warning: sun.rmi.server.UnicastServerRef is Sun proprietary API and may be removed in a future
> The JRMP exporter is just a way to use Exporters with standard RMI rather than JERI.
 You could convert that into a separate distribution that was specific to the Sun/Oracle JVM.
 Of course, since the things in question are part of RMI, they are probably part of most Java
implementations anyway.  Definitely separate from JERI.

In this case SunJrmpExporter is unique to Phoenix, the exporter in 
net.jini.jrmp.JrmpExporter doesn't utilise private namespace.  
SunJrmpExporter allows a known java.rmi.server.ObjID to be used, while 
JrmpExporter doesn't.

The question is: "Does Phoenix need to use a Registry exported over 
JRMP, for activatable services that use a JRMP exporter?"

>> import sun.rmi.server.MarshalInputStream;
>> C:\Users\peter\Documents\NetBeansProjects\peterConcurrentPolicy\src\com\sun\jini\phoenix\RegistrySunExporter.java:29:
warning: sun.rmi.server.UnicastServerRef is Sun proprietary API and may be removed in a future
> Hmm, not as sure about this one.

> Is this exporter needed for Phoenix?


If we wanted activation, we'd need to replace the functionality in 
java.rmi.registry.LocateRegistry, to allow a Registry to be exported 
over JERI endpoints.  Registry is a service provided at a known 
location, used for bootstrapping.

>   Do you view activation as an important part of JERI, or could that be split out as

I guess it would depend on the amount of work required, the Activation 
interfaces could stay, the implementation, Phoenix, can't without first 
replacing functionality in classes that access sun private namespace , 
so it would be provided separately.

Thanks Tim, glad to see you're still monitoring the list ;)



View raw message