Win a copy of Microservices Testing (Live Project) this week in the Spring forum!

Nick Bauman

+ Follow
since Jan 03, 2004
Cows and Likes
Total received
In last 30 days
Total given
Total received
Received in last 30 days
Total given
Given in last 30 days
Forums and Threads
Scavenger Hunt
expand Ranch Hand Scavenger Hunt
expand Greenhorn Scavenger Hunt

Recent posts by Nick Bauman

You metion formal processes wrt ClearCase. ClearCase's best practices are very invasive. They essentially goad you to go down the path of Dynamic Repository, which means that when someone checks in, _everyone_ gets their update whether they want it or not. The stick they use is the fact that there is no built-in way to recursively checkout a directory of resources: you have to write programs to do that for you. And did I mention ClearCase is very sloooowwwwww?

I wish I could advocate for Subversion, but I'm being cut off at every turn.

We are in the process of being forced to upgrade from CVS to ClearCase because of a management decision made 3 years ago. The ClearCase system which has been forced down other groups' throats is constantly going down (like every other day lately, but usually at least once a week).
The key phrase you are missing is "in the middle of a business method invocation". It is true that while a StatelessSessionBean is in the middle of servicing a request, it is tied to that client.
Kathy and Bert,
If I take the Business Interface approach (discussed on page 86 HFE), I can reap the benefits of using refactoring tools like those found in IntelliJ and Eclipse, even with CMP 2.0 Entity Beans, which must be declared abstract.
The only fly in the ointment that I'm finding is that this Business Interface is an all-or-nothing thing in terms of Local vs. Remote. The Business Interface must declare a throws RemoteException or not, depending on whether it has a Local or Remote API. Am I missing something or is there a nice way to get the type safety and flexibility on in-container vs out-of-container clientele?
Sounds like a really bad idea, as it breaks MVC and "Seperation of Concerns" principle. Not to mention breaking the Law of Demeter.
Probably a dumb question.
When I make an Entity Bean and deploy it using the RI, the Verifier complains that I have a transaction attribute on getEJBMetaData, but when I try to disable any transaction attributes in the GUI, I can only pick one of three: Required, Requires New and Mandatory. I can't disable and I can't pick any other Transaction Attrib.
Thanks for any insights
I have the book. I bought it a long time ago, when CMP 2.0 just came out. I've read it cover to cover (which is why HFEJB can be a little slow at times, no offense). But this was a long time ago, and I didn't have any EJB projects to work on at the time so it and my memory languished.
I haven't taken the exam, but I will say this about the book Mastering EJB: it has very few real app server examples. And the few examples it does have are for BEA Weblogic. Dealing with BEA's expireware server is not fun, either, unless you're unemployed and plan to scream through the book in a couple of weeks. Therefore, getting the examples to work with, say JBoss, is quite a trying experience.
Other than that, it's a really good book. Covers everything.
The introduction of the Remote vs. Local parlance in the EJB 2.0 spec is a great step forward in developing scalable applications using J2EE. It would be nice to use the same syntax for interacting with both types of EJBObjects for the same bean, modifying whether we work with a distributed stub or a local ref via a discriptor or some other type of runtime / deploytime switchability. Does anyone know whether this is forthcoming anytime soon?
I'm only on page 80 in HFEJB and I'm finding that reading the spec is unavoidable for questions like "What features must an EJB 2.0 compliant container have?".
I have been reading that Hibernate can be used as a BMP solution.
What I don't understand is that CLOBs are supposed to be supported in the JDBC spec for quite some time as a standard type (like, I think, as a String). I remember when PostgreSQL had trouble with JDBC standard large object support until they finally incorporated their large object support in their JDBC 2.0 driver over a year ago. So why wouldn't this work with Oracle today?
Just my thinking.
Hello, I'm replying to myself. Hehe. I found the answer elsewhere on this BBS.
CORBA problem resolved
(by the way, why is JavaRanch running a BBS that's a CGI program??!!)
From my old C++ days, you should always write a no-arg constructor of your own because the compiler-provided constructor may provide "special" behavior that you are unintentially relying on which might go away when you do one day write your own constructor.
Not sure this means the same thing in Java, since I don't write Java compilers, but it seems on-target because the black-box of a container is relying on something you've nailed down at compile-time.
I feel really dumb. I tried to do the simple brain-dead example in the HFEJB book, the AdviceGuy and I keep getting a CORBA exception when running the client. I followed the examples in the book like a parrot (after trying to do it on my own a couple of times) I'm using JDK 1.4.2, could that be the problem? What could be causing this?
java.rmi.RemoteException: CORBA BAD_OPERATION 0 No; nested exception is:
org.omg.CORBA.BAD_OPERATION: vmcid: 0x0 minor code: 0 completed: No
at javax.rmi.CORBA.Util.mapSystemException(
at headfirst._Advice_Stub.getAdvice(Unknown Source)
at AdviceClient.go(
at AdviceClient.main(
Caused by: org.omg.CORBA.BAD_OPERATION: vmcid: 0x0 minor code: 0 completed: No
at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
at sun.reflect.NativeConstructorAccessorImpl.newInstance(
at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(
at java.lang.reflect.Constructor.newInstance(
at java.lang.Class.newInstance0(
at java.lang.Class.newInstance(
at org.omg.CORBA.portable.ObjectImpl._invoke(
at headfirst._Advice_Stub.getAdvice(Unknown Source)
... 2 more
[ January 03, 2004: Message edited by: Nick Bauman ]