Java RMI questions and MyEclipse

[Clayton Cramer]

I hope that I am asking this in the right group.

1. I am looking at the possibility of replacing our browser-based JSP/Struts client talking to Java server classes with a thick client written in Java.. (Yes, for most enterprise systems, this would not be a great idea, but for us it does make sense: we have a few hundred PCs talking to our servers;there are never one time users connecting to our servers unless it's a security breach.)

2. I I started with the Oracle demo where the client makes a request for the value of pi to an arbitrary number of places, and the server computes it and passes it back to the client.

I was able to get this working using javac and jar from the command line (which involves bundling the class files into a JAR, copying it over to the tomcat webapps directory). My guess is that there is some way to set this upin MyEclipse so that it compiles and redeploys--but it is not exactly obvious how to set this up.

I am using MyEclipse 9.1, Tomcat 5.0.28 (please don't laugh; we're working on moving into modern times, and getting rid of our papyrus and clay tablets).

To simplify figuring out what is broken, I have taken the Oracle tutorial code and put in a single project, divided into three packages: client, compute, and engine. I can start the server (engine) with the following DOS batch file:

java -cp "c:\Documents and Settings\ccramer\Workspaces\MyEclipse 9\rmi\src";c:\tomcat5\webapps\rmi\compute.jar -Djava.rmi.server.codebase=file:/c:\tomcat5\webapps\rmi\compute.jar -Djava.security.policy=server.policy engine.ComputeEngine

I can run the client side from the command line with the following DOS batch file:

view plaincopy to clipboardprint?
java -cp "c:\Documents and Settings\ccramer\Workspaces\MyEclipse 9\rmi\src";c:\tomcat5\webapps\rmi\compute.jar -Djava.rmi.server.codebase=file:/c:\tomcat5\webapps\rmi\computer.jar -Djava.security.policy=client.policy client.ComputePi localhost %1%


where %1% is the parameter passed to the batch file specifying the number of digits of pi to have the server side calculate and return.

Then I tried to get MyEclipse to start the server in pretty much the equivalent way using the Run Configurations...:

Main.class: engine.ComputeEngine
Classpath: default classpath and c:\tomcat5\webapps\rmi\compute.jar
Environment: java.rmi.server.codebase=file:/c:\tomcat5\webapps\rmi\compute.jar
java.security.policy=server.policy

Code:
ComputeEngine exception:
java.rmi.ServerException: RemoteException occurred in server thread; nestedexception is:
java.rmi.UnmarshalException: error unmarshalling arguments; nested exception is:
java.lang.ClassNotFoundException: compute.Compute
at sun.rmi.server.UnicastServerRef.oldDispatch(UnicastServerRef.java:385)
at sun.rmi.server.UnicastServerRef.dispatch(UnicastServerRef.java:240)
at sun.rmi.transport.Transport$1.run(Transport.java:153)
at java.security.AccessController.doPrivileged(Native Method)
at sun.rmi.transport.Transport.serviceCall(Transport.java:149)
at sun.rmi.transport.tcp.TCPTransport.handleMessages(TCPTransport.java:460)
at sun.rmi.transport.tcp.TCPTransport$ConnectionHandler.run(TCPTransport..java:701)
at java.lang.Thread.run(Thread.java:595)
at sun.rmi.transport.StreamRemoteCall.exceptionReceivedFromServer(StreamRemoteCall.java:247)
at sun.rmi.transport.StreamRemoteCall.executeCall(StreamRemoteCall.java:223)
at sun.rmi.server.UnicastRef.invoke(UnicastRef.java:343)
at sun.rmi.registry.RegistryImpl_Stub.rebind(Unknown Source)
at engine.ComputeEngine.main(ComputeEngine.java:29)
Caused by: java.rmi.UnmarshalException: error unmarshalling arguments; nested exception is:
java.lang.ClassNotFoundException: compute.Compute
at sun.rmi.registry.RegistryImpl_Skel.dispatch(Unknown Source)
at sun.rmi.server.UnicastServerRef.oldDispatch(UnicastServerRef.java:375)
at sun.rmi.server.UnicastServerRef.dispatch(UnicastServerRef.java:240)
at sun.rmi.transport.Transport$1.run(Transport.java:153)
at java.security.AccessController.doPrivileged(Native Method)
at sun.rmi.transport.Transport.serviceCall(Transport.java:149)
at sun.rmi.transport.tcp.TCPTransport.handleMessages(TCPTransport.java:460)
at sun.rmi.transport.tcp.TCPTransport$ConnectionHandler.run(TCPTransport..java:701)
at java.lang.Thread.run(Thread.java:595)
Caused by: java.lang.ClassNotFoundException: compute.Compute
at java.net.URLClassLoader$1.run(URLClassLoader.java:200)
at java.security.AccessController.doPrivileged(Native Method)
at java.net.URLClassLoader.findClass(URLClassLoader.java:188)
at java.lang.ClassLoader.loadClass(ClassLoader.java:306)
at java.lang.ClassLoader.loadClass(ClassLoader.java:251)
at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:319)
at java.lang.Class.forName0(Native Method)
at java.lang.Class.forName(Class.java:242)
at sun.rmi.server.LoaderHandler.loadProxyInterfaces(LoaderHandler.java:707)
at sun.rmi.server.LoaderHandler.loadProxyClass(LoaderHandler.java:651)
at sun.rmi.server.LoaderHandler.loadProxyClass(LoaderHandler.java:588)
at java.rmi.server.RMIClassLoader$2.loadProxyClass(RMIClassLoader.java:628)
at java.rmi.server.RMIClassLoader.loadProxyClass(RMIClassLoader.java:294)
at sun.rmi.server.MarshalInputStream.resolveProxyClass(MarshalInputStream.java:238)
at java.ibjectInputStream.readProxyDesc(ObjectInputStream.java:1494)
at java.ibjectInputStream.readClassDesc(ObjectInputStream.java:1457)
at java.ibjectInputStream.readOrdinaryObject(ObjectInputStream.java:1693)
at java.ibjectInputStream.readObject0(ObjectInputStream.java:1299)
at java.ibjectInputStream.readObject(ObjectInputStream.java:339)
... 9 more


As you can see, there is a ClassNotFoundException: compute.Compute

I'm hard pressed to see why. If I do a jar -tf c:\tomcat5\webapps\rmi\compute.jar I get:

Code:
META-INF/
META-INF/MANIFEST.MF
compute/Compute.class
compute/Task.class


Any suggestions?

 

[Arne Vajhøj]

I hope that I am asking this in the right group.

1. I am looking at the possibility of replacing our browser-based
JSP/Struts client talking to Java server classes with a thick client
written in Java. (Yes, for most enterprise systems, this would not
be a great idea, but for us it does make sense: we have a few hundred
PCs talking to our servers; there are never one time users connecting
to our servers unless it's a security breach.)

2. I I started with the Oracle demo where the client makes a request
for the value of pi to an arbitrary number of places, and the server
computes it and passes it back to the client.

I was able to get this working using javac and jar from the command
line (which involves bundling the class files into a JAR, copying it
over to the tomcat webapps directory). My guess is that there is some
way to set this up in MyEclipse so that it compiles and
redeploys--but it is not exactly obvious how to set this up.

I am using MyEclipse 9.1, Tomcat 5.0.28 (please don't laugh; we're
working on moving into modern times, and getting rid of our papyrus
and clay tablets).

I would recommend against using RMI here. A web container is not a
good RMI server.

Expose the functionality as web services and let your fat client
call those.

To simplify figuring out what is broken, I have taken the Oracle
tutorial code and put in a single project, divided into three
packages: client, compute, and engine. I can start the server
(engine) with the following DOS batch file:

java -cp "c:\Documents and Settings\ccramer\Workspaces\MyEclipse
9\rmi\src";c:\tomcat5\webapps\rmi\compute.jar
-Djava.rmi.server.codebase=file:/c:\tomcat5\webapps\rmi\compute.jar
-Djava.security.policy=server.policy engine.ComputeEngine

I can run the client side from the command line with the following
DOS batch file:

view plaincopy to clipboardprint? java -cp "c:\Documents and
Settings\ccramer\Workspaces\MyEclipse
9\rmi\src";c:\tomcat5\webapps\rmi\compute.jar
-Djava.rmi.server.codebase=file:/c:\tomcat5\webapps\rmi\computer.jar
-Djava.security.policy=client.policy client.ComputePi localhost %1%

where %1% is the parameter passed to the batch file specifying the
number of digits of pi to have the server side calculate and return.

Then I tried to get MyEclipse to start the server in pretty much the
equivalent way using the Run Configurations...:

Main.class: engine.ComputeEngine Classpath: default classpath and
c:\tomcat5\webapps\rmi\compute.jar Environment:
java.rmi.server.codebase=file:/c:\tomcat5\webapps\rmi\compute.jar
java.security.policy=server.policy

Code: ComputeEngine exception: java.rmi.ServerException:
RemoteException occurred in server thread; nested exception is:
java.rmi.UnmarshalException: error unmarshalling arguments; nested
exception is: java.lang.ClassNotFoundException: compute.Compute at
sun.rmi.server.UnicastServerRef.oldDispatch(UnicastServerRef.java:385)

As you can see, there is a ClassNotFoundException: compute.Compute

I'm hard pressed to see why. If I do a jar -tf
c:\tomcat5\webapps\rmi\compute.jar I get:

Code: META-INF/ META-INF/MANIFEST.MF compute/Compute.class
compute/Task.class

Can you check the actual java command used when running inside
Eclipse?

Arne

 

[Roedy Green]

1. I am looking at the possibility of replacing our browser-based JSP/Strut=
s client talking to Java server classes with a thick client written in Java=
. (Yes, for most enterprise systems, this would not be a great idea, but f=
or us it does make sense: we have a few hundred PCs talking to our servers;=
there are never one time users connecting to our servers unless it's a sec=
urity breach.)

I reject of thin clients on many grounds:
1. inefficiency.
2. bugs introduced by the browser
3. you don't keystroke by keystroke validation
4. there is no logic to help you fill out a form when not all the
fields are mandatory.

The problem with RMI is you expose everything to the outside world,
hopefully just an internal LAN.

I did two projects of this sort. The server had a socket open to each
terminal. The protocol was a continuous stream of large binary
packets (most commonly image files). It had quite spectacular
throughput.

 

[Arne Vajhøj]

I reject of thin clients on many grounds:
1. inefficiency.

And you are sure that it a problem because?

2. bugs introduced by the browser

Does browsers have more bugs than other environments?

Swing may be relative stable, but JavaFX probably have a few oopses.

3. you don't keystroke by keystroke validation

Ever heard about AJAX?

4. there is no logic to help you fill out a form when not all the
fields are mandatory.

See above.

Arne

 

[Joerg Meier]

I reject of thin clients on many grounds:
1. inefficiency.

In what sense ?

2. bugs introduced by the browser

As others said, that seems to be not much more of an issue than bugs
introduced by Swing or whatever.

3. you don't keystroke by keystroke validation
4. there is no logic to help you fill out a form when not all the
fields are mandatory.

This kind of sounds like you are stuck in ca. 2000 web technology. Are
there even still websites today for which the above is true, let alone
future websites ?

Liebe Gruesse,
Joerg

 

[Clayton Cramer]

Can you check the actual java command used when running inside

Eclipse?



Arne

I am not sure how to see what MyEclipse actually produces. I suspect that there is some magic menu choice that shows this, but I am not sure what it is.

 

[Lew]

I am not sure how to see what MyEclipse actually produces.
I suspect that there is some magic menu choice that shows this, but I am not sure what it is.

http://help.eclipse.org/juno/topic/org.eclipse.jdt.doc.user/gettingStarted/qs-BasicTutorial.htm

 

[Jeff Higgins]

I am not sure how to see what MyEclipse actually produces. I suspect that there is some magic menu choice that shows this, but I am not sure what it is.

Is this posting in any related to this two year old post at the
MyEclipse Forums?
<http://www.myeclipseide.com/PNphpBB2-viewtopic-t-29372.html>
Just curious, my grasp of this technology isn't great enough for me to
determine it.

 

[Jeff Higgins]

Is this posting in any related to this two year old post at the
MyEclipse Forums?

Oops.I should have slept a little longer.
Make that several days rather than two years.