Running SPARK from CVS comes with some extra requirements:
- A CVS client
- An account that can access cvs.ai.sri.com
- A Java SDK (must be 1.5.0 or later)
- Ant (Java-based build tool)
- Jython 2.2a1x: This is a special build of Jython that fixes several bugs. You can get a copy in spark/lib/jython22a1x.zip. To install, simply create a directory (e.g. 'jython') and unzip the contents of jython22a1x.zip into it.
Windows/Unix/Mac OS X (cvs.ai.sri.com tree only):
- Checkout the 'spark' module from CVS (server: cvs.ai.sri.com, cvsroot=/homedir/cvs/cvsroot, ext protocol)
- Compile spark using ant (type 'ant' in the root directory of the module)
- Set the JYTHON_HOME location (e.g. c:\jython\). You have two options for setting this. You can either define it as an environment variable (on Windows, use Control Panel->System->Advanced->Environment Variables), or you can edit spark/bin/spark.bat (or on Unix/Mac OS X edit spark/bin/spark) and specify the Jython location where noted in the file.
- Set the SPARK_MODULES location. For more information on this, please see "The spark_modules directory" and "Setting the location of your SPARK modules" in the Running SPARK documentaiton.
- Optional: it is recommended that you add spark/bin to your system path.
You should now be ready to start running SPARK.
NOTE: If you are going to be using SPARK with the CALO Desktop, you may wish to read the CALO Desktop notes in the End-User installation instructions.