[jira] [Created] (STRATOS-110) Topology Event Message Processors should follow 'chain of responsibility' design pattern

[Nirmal Fernando (JIRA)]

Nirmal Fernando created STRATOS-110:
---------------------------------------

             Summary: Topology Event Message Processors should follow 'chain of 
responsibility' design pattern
                 Key: STRATOS-110
                 URL: https://issues.apache.org/jira/browse/STRATOS-110
             Project: Stratos
          Issue Type: Improvement
    Affects Versions: 4.0.0 M1
            Reporter: Nirmal Fernando
            Assignee: Nirmal Fernando
            Priority: Critical
             Fix For: 4.0.0 M1


Following is the MessageProcessor interface with this change.

package org.apache.stratos.messaging.message.processor;

import org.apache.stratos.messaging.domain.topology.Topology;

/**
 * Message processor interface. Every Message Processor should implement this.
 */
public interface MessageProcessor {
   
    /**
     * Link a message processor and its successor, if there's any.
     * @param nextProcessor
     */
    public abstract void setNext(MessageProcessor nextProcessor);

    /**
     * Message processing and delegating logic.
     * @param type type of the message.
     * @param message real message body.
     * @param topology Topology that will get updated.
     * @return whether the processing was successful or not.
     */
    public abstract boolean process(String type, String message, Topology 
topology);
}

I've implemented initial Message Processors:

├── ClusterCreatedEventProcessor.java
├── ClusterRemovedEventProcessor.java
├── CompleteTopologyEventProcessor.java
├── MemberActivatedEventProcessor.java
├── MemberStartedEventProcessor.java
├── MemberSuspendedEventProcessor.java
├── MemberTerminatedEventProcessor.java
├── ServiceCreatedEventProcessor.java
└── ServiceRemovedEventProcessor.java

Now, it is up to the Message Processor Delegator to build the Message processor 
chain.

        // instantiate all the relevant processors
        ServiceCreatedEventProcessor processor1 = new 
ServiceCreatedEventProcessor();
        ServiceRemovedEventProcessor processor2 = new 
ServiceRemovedEventProcessor();
        ClusterCreatedEventProcessor processor3 = new 
ClusterCreatedEventProcessor();
      
       
        // link all the relevant processors in the required order
        processor1.setNext(processor2);
        processor2.setNext(processor3);
        processor3.setNext(processor4);

and also it's a duty of the deligator to start the flow:

boolean hasProcessed = processor1.process(type, json, 
TopologyManager.getTopology());

After the flow started, Processor who is capable of handling the message would 
eventually receive it and will process the message and other Processors will 
simply delegate the message to its successor.




--
This message was sent by Atlassian JIRA
(v6.1#6144)