2 Replies Latest reply on Feb 25, 2006 12:46 PM by Ronald van Kuijk

    3 ways to transition: processInstance.signal, token.signal,

    jbpmndc Newbie

      There are 3 ways to transition to another node:
      * processInstance.signal("somewhere")
      * token.signal("somewhere")
      * taskInstance.end("somewhere")

      Why so many ways?

      I like taskInstance.end("somewhere"), because it ends the task. The signal just moves from one to another. I don't believe it "ends" a task.

      Any thoughts?

        • 1. Re: 3 ways to transition: processInstance.signal, token.sign
          prabuddha Newbie

          I've the same opinion as you and I'm facing a problem in invoking the assignment handler through taskinstance.end(). Its that while token.signal() invokes the assignment handler class taskinstance .end does not.

          can any one explain me.

          • 2. Re: 3 ways to transition: processInstance.signal, token.sign
            Ronald van Kuijk Master

            Why should ending a task activate an assignmenthandler? The task is finished and noone has to act upon it. Ending a task does not automatically mean that the tasknode is completed. It could be that there are more tasks in the same tasknode. Depending on the configuration of the tasknode (should it wait for all tasks to finish, or continue if one is finished) the tasknode is completed (token internally signalled) and progressed to the next node. (it is all in the javadoc on the code...)

            That next node can have an assignmenthandler which will be called then.

            So ending a task, signalling a specific token and signalling the processInstance (which signals the token of the current node according to the comments, but it signals the rootToken) all do different things.

            afaik, the docs are clear on this, but if they aren't, please point us to the ommissions.