I've been looking in the docs for jBPM 6.1 Final and here on the forums for information regarding service repositories and have some questions about them.
I created a repository with a work item definition and 2 jar files listed as dependencies:
I was able to import the work item definition and have it added to the service palette in the jbpm-console business process editor.
1. What exactly has this achieved?
I ask because I attempted to create a process variable of a type defined in the datamodel.jar but that fails since the console doesn't seem to have added that jar to the classpath.
If I skip creating the process variable and just try to invoke the custom work item handler, I get an error in the log saying the console can't find a work item handler of the type that was specified in the work item definition file.
2. Do I still need to edit the CustomWorkItemHandlers.conf file and register an instance of the new custom work item handler?
3. Is the service repository import code expecting that one of the dependency jars is a full-fledged kjar? In other words, if the customhandler.jar from this example had a filled out kmodule.xml file, would it work? My test one has a kmodule.xml but the kmodule tag inside doesn't have any child elements defined.
4. After I've imported a service, is there a way to delete it? If I import a service that is broken or old or unused, how do i remove it from the palette? If the service gets updated, how do i "refresh" the version that was previously imported?
I get the basics of what the service repository is trying to achieve but it feels like I'm missing something or there's still some work to do in terms of documenting its usage.
Let's simplify my post a little bit and keep it to one question at a time.
If I connect to a service repository and the .wid for a particular service lists a couple jars as dependencies, after I install that service, what has happened to those jars? I would expect them to be deployed in my jbpm-console artifact repository... Is that the expected behavior?