Because we want our Command to include Java bytecode, we use RMI rather than a web service. Not that we couldn’t use a web service at all, but it’s just easier with RMI. If you’re totally out of the loop on RMI, please check out my free RMI tutorial.
This repo has two separate projects, server and client, each configured both for Maven and Eclipse. After you’ve cloned this repo, here’s how to get it going. Instructions are for *Nix-based systems; for MS-DOS-based systems you’ll have to do a bit of translation of commands and shell scripts, but it ain’t rocket science.
Note that the "common" code is in the server, so the server class files have to be available to compile the client. Hence, the order you do things in matters. In Maven we install the client code into your personal Maven repository. In Eclipse we make the client project depend on the server project.
cd server; mvn package install // client depends on server code ./runserver & // script looks after running rmiregistry cd .. cd client; mvn package ./runclient
Create server project
-
File→New Java Project
-
Click OFF "Use default location".
-
Set name to "remotecommand-server" (this name’s important as the client project depends on it)
-
Navigate to git repo, single click on server, click Open button.
-
Click Finish.
Create client project
-
File→New Java Project
-
Click OFF "Use default location".
-
Set name to "remotecommand-client" (this name’s less important but why not keep these two projects together in your projects list)
-
Navigate to git repo, single click on client, click Open button.
-
Click Finish.
You can then manually run the RMI registry in another window, run the server’s main program, and run the client’s main program, and see the interaction.