JBPM3中文用户手册——第三章：教程(4)

3.1    任务分配示例

下一个例子，我们将展示给你如何分配任务给用户。由于jbpm工作流引擎与组织模型是分开的，仅有用于计算用户标识号的表达式语言经常是不够的。因此，你必须指定一个AssignmentHandler接口的实现，其中包含对任务参与者的计算。

 

public void testTaskAssignment() {   // The process shown below is based on the hello world process.   // The state node is replaced by a task-node.  The task-node   // is a node in JPDL that represents a wait state and generates   // task(s) to be completed before the process can continue to   // execute.    ProcessDefinition processDefinition = ProcessDefinition.parseXmlString(     "<process-definition name='the baby process'>" +     "  <start-state>" +     "    <transition name='baby cries' to='t' />" +     "  </start-state>" +     "  <task-node name='t'>" +     "    <task name='change nappy'>" +     "      <assignment class='org.jbpm.tutorial.taskmgmt.NappyAssignmentHandler' />" +     "    </task>" +     "    <transition to='end' />" +     "  </task-node>" +     "  <end-state name='end' />" +     "</process-definition>"   );     // Create an execution of the process definition.   ProcessInstance processInstance =       new ProcessInstance(processDefinition);   Token token = processInstance.getRootToken();     // Let's start the process execution, leaving the start-state   // over its default transition.   token.signal();   // The signal method will block until the process execution   // enters a wait state.   In this case, that is the task-node.   assertSame(processDefinition.getNode("t"), token.getNode());     // When execution arrived in the task-node, a task 'change nappy'   // was created and the NappyAssignmentHandler was called to determine   // to whom the task should be assigned.  The NappyAssignmentHandler   // returned 'papa'.     // In a real environment, the tasks would be fetched from the   // database with the methods in the org.jbpm.db.TaskMgmtSession.   // Since we don't want to include the persistence complexity in   // this example, we just take the first task-instance of this   // process instance (we know there is only one in this test   // scenario).   TaskInstance taskInstance = (TaskInstance)        processInstance         .getTaskMgmtInstance()         .getTaskInstances()         .iterator().next();     // Now, we check if the taskInstance was actually assigned to 'papa'.   assertEquals("papa", taskInstance.getActorId() );     // Now we suppose that 'papa' has done his duties and mark the task   // as done.   taskInstance.end();   // Since this was the last (only) task to do, the completion of this   // task triggered the continuation of the process instance execution.     assertSame(processDefinition.getNode("end"), token.getNode()); }