Chris Bicnal

I'm trying to implement a LogicalHandler in my Axis2 client with the goal of filtering out some problematic characters (originating from a mainframe) present in the message payload which are causing an issue with Axis2 parsing the XML into a SOAPMessage.

I have created the LogicalHandler and have annotated the generated service class with @HandlerChain(file="../../axis2/handlers/handler.xml") referencing the following;

<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <javaee:handler-chains xmlns:javaee="http://java.sun.com/xml/ns/javaee" xmlns:xsd="http://www.w3.org/2001/XMLSchema"> <javaee:handler-chain>  <javaee:handler>   <javaee:handler-class>com.bic.axis2.handlers.InvalidCharHandler</javaee:handler-class>  </javaee:handler> </javaee:handler-chain> </javaee:handler-chains>


...the path is correct but when I fire the web service my handler isn't invoked.

What am I doing wrong?

Bic

Anton,

How about we use both?

Store the method name > class mappings in a Map and then use reflection to instantiate. If they all inherit from the same parent my delegation logic is very short and quick - and thread safe.

I can populate the Map from a configuration file or data source once on initiation, but it lends itself to growth pretty well.

What do you think?

Chris

Anton,

The centralisation did spring to mind when thinking about this - logging/auditing, authentication etc.. would be simplified.

I think my bigger question boils down to, what's the best way to implement an efficient mechanism that accepts string method names and parameters and translates them into actually method calls.

Bic

Hey guys,

I'm trying to implement an interface from a mobile application and am wondering if the Front Controller pattern is the correct way to do this.

The interface I have to the mobile application is a single java method;

public String sendAndReceive(String methodName, String parameters)

...the parameters are passed in as a JSON object.

The goal is for me to take the method name and JSON parameters, translate them into a call to a back end system and translate the return into a JSON object for consumption by the mobile application.

There are roughly 60 different method calls that the mobile application could make. I need to figure out the best way to invoke the appropriate logic as was thinking the FrontController pattern would be best placed for this - using a dispatcher to do the mapping of methodName to classes which would manage the invocation of the method.

My concern is that I'm going to have 60 different classes (and a big dispatcher method with lots of if else constructs!) if I do this.

Any thoughts?

Bic

Naren,

I purposely wanted to do it quickly so it wasn't hanging over my head.

I started working on it during my Christmas vacation and probably spent 20 hours or so then. Once back at work after the break I was working on it during evenings and weekends. All in all I estimate I put in another 20-25 hours over the course of the three weeks in January leading up to my submission and exam. It was probably 40-50 hours of effort all in.

It seems like a daunting task at first but I broke it down and tackled...

1. Class Diagram
2. Sequence Diagrams

...once those two are done you've 70% there IMO.

The class diagram helps frame your solution and the sequence diagrams, although only worth a few points, really help to solidify your design and validate your class diagram - so I recommend putting in the effort here.

3. Component Diagram
4. Deployment Diagram

The component diagram is basically an abstracted view of your class diagram (didn't take me long) and the deployment diagram is as simple or complex as you chose to make it.

5. Risks/Assumptions

I bought myself a note pad specifically for the project and kept all my notes and assumptions scribbled there - so when it came to #5 it was really a case of reading through and consolidating my thoughts. I recommend this approach - it's very useful to have something you can carry around to jot thoughts down if you're on the subway, at home or in the office.

I also recommend getting your diagrams down on paper before trying to write them up in a tool. The notebook is useful for this as you go through various iterations of your diagrams you're able to look back and see the history of your thinking. The old fashioned approach also enables you to focus on exactly what the problem is as opposed to being distracted by the tool (which is especially true if you're new to the tool you've decided to use).

I hope that helps!

Chris

Hi all,

I received notification today that I cleared part II and III with a score of 146!

I had the Gemstone International project and am very happy that my work over the Christmas vacation has all paid off! :-)

Thanks to all on this forum for their help, but a special mention to K Tsang for all the guidance and advice provided.

In terms of timelines, I purchased the assignment on Dec 22nd 2015, submitted on Jan 22nd, took the exam Jan 23rd and received my result today.

This certification has taught me a great deal and I'm pleased to say that I find myself using the knowledge gained in my day to day work.

Thanks again ranchers!

Chris

PS Happy to answer any questions or provide advice on the process!

My question was more about a) being lazy and b) not wanting to produce messy HTML that may have issues across platforms and browsers.

Producing the deliverables is more akin to word processing than web development, so I didn't want to create Word HTML or write my document in Eclipse.

In the end I chose KompoZer which is a really neat little WYSIWYG editor that doesn't create excess or messy HTML.

Cheers,

Chris

I feel a little silly for asking this question, but how did folks write the HTML index file for submission?

Did you write it yourself or use a WYSIWYG editor? If so, which one produces the cleanest HTML?

Cheers,

Chris

Thanks Jeanne - can I assume from your reply that you also included your risk list and assumptions within the index.html document as opposed to linking to separate files?

Howdy all,

So I'm in the final stages of packaging up my application and have a question on what folks have done about comments.

I have a ton of comments on my deployment diagram and am wondering whether it's better to leave them on the diagram or move them to index.html with a link to the "clean" diagram from there.

What did folks do?

Chris

Hi there,

Am I the only one NOT to use JMS in my assignment?

Did anyone else not use it and pass?

Chris

I have a question into Oracle Education - they have acknowledged and will get back to me once they've completed their research (eta 3 business days).

Thanks L - I had someone on the Ranch take a look and got great feedback.

If anyone else would like to offer their help then feel free to Purple Mooseage me and I'll be in touch!

Thanks for the info K - very useful as ever!

I like the Timer Bean approach as, when I have a clustered application server, I'm not dependent on any one single node. This availability is harder to create when using a server farm, cron job or other mechanism.....unless I'm missing something??

As to Oracle expiring the certification, I have until June as per my exam so I'm wondering whether this is just for Part 1 and anyone who wants to start Part 2 after March? Either way - who/how do I contact for more information on this?

Chris

Hi all,

I'm working on part II of the JEE5 stream and am using timer beans.

The thing is, JEE6 introduced calendar based scheduling, which is a really nice addition that would help me - and simplify my life!

Even though I'm on the JEE5 track would I be penalised for including items from JEE6?

Has anyone done this?

Thanks,

Chris