Paulo Salgado

I was reading about the subject above on the epractizelabs web site and some questions popped up.

During the Plan phase, regarding the scheduling of achieving the goals, how exactly is it organized? By objective, section, is it selectable?

When setting the schedule is there any sort of estimate of how long to complete each task or is it estimated by the user?

Are the goals set by a percentage of right answers on the Practize phase?

Are there any references for the user when the tool identifies an objective that should be given more focus?

Thanks and regards.

-Paulo

Hi Aditi.

I'm not a Weblogic guy, but I was looking over the 8.1 docs and it seems it implements J2EE 1.3 and JAX-RPC 1.0. Being that the case, the method of deployment of Message Handlers is not mandated by the spec. See JAX-RPC 1.0 12.3.2. That changes for J2EE 1.4 when webservices.xml is used.

There seems to be some good docs for Weblogic though. Take a ook at this one, which guides you specifically on designing, building and deploying Message Handlers with Weblogic 8.1: "Creating SOAP Message Handlers to Intercept the SOAP Message"

And this one is the parent doc. Lots of info.

Hope it helps and don't forget to share the results of your implementation.

Regards.

Hi Marzio.

If you want to create documentation so the users of your Web Services can create client code to use it, you only need to create a WSDL file. If you don't want to generate a WSDL file from scratch, some SOAP engines provide a tool called Java2WSDL, which will create a WSDL file from a Java class.

If you want to provide verbal documentation about your WS then I still think it would go into the WSDL, but I haven't done that. I'm also not aware of a Javadoc-like tool for WSDL files, but there may be one. Hopefully somebody else has done it and will post the experience.

Regards.

Hi Aditi.

Considering you will be using JAX-RPC (or the new JAX-WS) or even Apache Axis, one way of logging your requests and responses is by implementing a Message Handler for that purpose. A Message Handler acts like a Servlet Filter, i.e., it intercepts your request before it reaches its destination (the endpoint) and intercepts the response before it is returned to the caller (the WS client). Both are very good times to log what is going back and forth.

You can write your Message Handler to use your favorite logging API, get the XML contents and post them to a log file.

I personally would recommend you to create a JAX-RPC (or JAX-WS) compliant Message Handler because that would make it reusable to any JAX-RPC compliant SOAP engine. And that means you could potentially use the same logging Message Handler on the client, on the server and on other applications. All you have to do is deploy it to the desired Web Service once it's complete.

Regards.

Hi Sanat.

It looks like you have a conflict between the element name and the type name. JAXB will generate interfaces and classes for both, so they can't have the same name. Try renaming either one of them. For instance, rename the element on your XML Schema like this:

<xs:schema xmlns="urn:example.com/test" xmlns:impl="urn:example.com/test" xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns:jxb="http://java.sun.com/xml/ns/jaxb" xmlns:xjc="http://java.sun.com/xml/ns/jaxb/xjc" targetNamespace="urn:example.com/test" elementFormDefault="qualified" jxb:extensionBindingPrefixes="xjc" jxb:version="1.0"> <xs:element name="[B]addRequestElement[/B]" type="addRequest" /> <xs:complexType name="addRequest"> <xs:sequence> <xs:element name="n1" type="xs:int"></xs:element> <xs:element name="n2" type="xs:int"></xs:element> </xs:sequence> </xs:complexType> </xs:schema>

Regards.

I second that, specially if you're thinking about developing Web Services with Java, as data binding has been removed from the new JAX-RPC (JAX-WS). J2EE vendors will need to provide or use a JAXB implementation to be compliant.

As a feedback we've been using JAXB 1.1 for over a year without problems. Our XML Schemas are somewhat complex and JAXB handles them nicely, something that WSDL2Java from both Apache Axis and IBM Websphere didn't.

Regards.

In theory you could create the UserTransaction before the EJB is called and define the EJB method with the "Supports" or "Required" transaction attribute. The EJB would use the same transaction context of the UserTransaction, which would also be available to Hibernate.
[ July 14, 2005: Message edited by: Paulo Salgado ]

Hi.
In fact my goal is to have a few versions/distributions in one hard drive. I'm definetely not planning to have multiple distributions in a single partition (at least for now), but the idea of having /home and/or /opt available for all of them sounds very interesting.
Thank you both for the answers and such a great advice.

Howdy.
I'm fairly new to the Linux/UNIX world and I'm trying to devise a partitioning scheme for multiple Linux installations in the same hard drive.
I've seen previous posts in this forum and got a bunch of good info as usual.
One question remains though: assuming I will have a single partition for each Linux version I will install, can I have a single Swap partition, or must I have a Swap partition for each corresponding Linux version partition?

I've read about this. Since then I started reading my mail offline. I often see query strings like id=76t327237 (whatever) in the URL, which probably links to your e-mail when they receive the request. Can't tell I was too successful though, as new spammers cme in every day... and my filters list grows...

One way to handle this kind of relationship is to add a table to your schema that contains the Primary Keys of both entities involved. For instance in your case you could have a table composed of PART_NUMBER (for parts) and MAKE + MODEL (for cars).
Whenever you need to relate both entities use this table in a join operation, so you can retrieve columns from each individual entity.
Regards.
-Paulo

is it possible that the AS400 server is denying connection to the database because of the number of open connections?. Each page has it's own Instance of st,cn,rs, etc.

It is indeed. Check with your System Programmer or DBA the maximum number of concurrent connections allowed and see if you're exceeding it.

Not sure if you are testing the Warning thing or if you want to check whether the row was really updated. In the latter case, Statement.executeUpdate(...) should return an int, in your case it should be zero and you could go from there.

Are you talking about Dataset Names?

Another option would be to use rs.getString("CREATE_DATE") instead of rs.getDate("CREATE_DATE"), as the TO_CHAR function probably returns a String.
-Paulo
[ July 25, 2003: Message edited by: Paulo Salgado ]
[ July 25, 2003: Message edited by: Paulo Salgado ]