At 14:00 1/11/00 +0000, Aaron M. Stromas wrote:
[...]
>bash$ java -Djava.security.policy=/home/ams/.java.security Client
>java.rmi.UnmarshalException: error unmarshalling return; nested
>exception is:
> java.lang.ClassNotFoundException: com.wrox.rmi.Reverse_Stub
>java.lang.ClassNotFoundException: com.wrox.rmi.Reverse_Stub
> at java.io.ObjectInputStream.inputObject(Compiled Code)
> at java.io.ObjectInputStream.readObject(Compiled Code)
> at java.io.ObjectInputStream.readObject(Compiled Code)
> at sun.rmi.registry.RegistryImpl_Stub.lookup(Compiled Code)
> at java.rmi.Naming.lookup(Compiled Code)
> at Client.main(Compiled Code)
RMIRegistry can't find the _Stub class. It needs three classes in order
to function (interface, _Stub, _Skel). The key to interpreting this from
the stack trace is to note that there are TWO exceptions nested, the outer
one is the UnmarshalException, but it is wrapping a ClassNotFound exception
which came accross the wire (io.ObjectInputStream.inputObject) from
the RMIRegistry on a lookup call. (RegistryImpl_Stub.lookup)
[EMAIL PROTECTED] apparently wrote in private email to Aaron:
> Don't the two vm's have to be the same, as the serial numbers will be
> different between 1.2 and 1.3, which will cause Serialization errors?
there can be serialization problems, but they're primarily between 1.1
and 1.2, or classes compiled by Sun's algorythm and another (i.e. Jikes)
because Sun used internal to javac constructs to generate the suuid.
At 12:25 1/11/00 -0800, Nathan Meyers wrote:
> RMI doesn't send classfiles, either for the real class or the stub.
it can be configured to however, not only the required _Stub/_Skel
gorp, but the entire dependency tree of classes. tends to make bringup
times really agonizing when you try to deploy on the net though. ;)
it looks like Aaron was/is on the way to doing this for the client,
but I'll doubt he even took a second thought about it when starting
the RMIRegisrty. (y'all have no idea how many times people have called
me to complain that my code is broken and it turned out they hadn't
restarted the server processes, or regenerated _Stub/_Skel(s) after
changing their own code; many of them turn shades of red/purple when
you deduce the problem ;)
cabbey at home dot net <*> http://members.home.net/cabbey
I want a binary interface to the brain!
Today's opto-mechanical digital interfaces are just too slow!
----------------------------------------------------------------------
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
