• Post Reply
  • Bookmark Topic Watch Topic
  • New Topic

remarks on Kushner's book IBM287

 
cyril vidal
Ranch Hand
Posts: 247
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
Hi,

I've just two other remarks on the book of Mr. Kushner.

1�) about transactions p.425
It's said that
J2EE application client components may also perform transaction, but they must access the transaction directly through JNDI as follows:
UserTransaction transaction = (UserTransaction)jndiContext.lookup("jta/usertransaction");


It's contradictoty with the J2EE1.3 spec 9.3 p.118 that states that
application clients are not required to have direct access to the transaction facilities of the J2EE platform. A J2EE product is not required to provide a JTA UserTransaction object for use by application clients
.
So, the solution provided by the spec: use of EJB's transaction facilities as well as JDBC's ones.

2�) One month ago, I've asked about the location and the way to get the client-resource.xmi file where the the resource references for application client are resolved. NO RESPONSE FROM ANYONE IN THIS FORUM, as well as nothing except of the mention of the file name in MrKushner's book p.417).
I've found the solution by reading the excellent redbook EJB2.0 DEvelopment with websphere Studio Application Developer (p.632).
- Run the clientConfig.bat file
- navigate to the exported ear
- expand the client.jar -> JMS Providers -> Websphere JMS Provider -> WASQueueeConnectionFactory
=> a file named client-resource.xmi is added to the client.jar file inside the EAR file.
=> we can extract this file and import this into the META-INF folder of the client project so that no error occurs during execution in WSAD.
It's a pity that nothing about the Application Client Resource Configuration Tool was ibn written in ibm 287's book.

By the way, this book can be useful although i prefer the EJB2.0 development redbook that is much more practical and dives much deeper in details

Regards,
Cyril.
[ February 07, 2005: Message edited by: cyril vidal ]
 
Axel Janssen
Ranch Hand
Posts: 2166
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
Cyril,

I don't see point 1 as a contradiction. You can understand the "not required " in EJB spec as optimal feature, but not require to be J2EE 1.3 compliant. As far as I see, IBM does understand it that way.
In real world starting transactions in client might create problems.
This is IBM test. On certain issues in J2EE there might exist s.th. like IBM POV.
Only god knows truth and truth is sometimes app specific.

regards Axel
[ February 08, 2005: Message edited by: Axel Janssen ]
 
cyril vidal
Ranch Hand
Posts: 247
  • Mark post as helpful
  • send pies
  • Quote
  • Report post to moderator
Hi Axel,

You're right, it's quite ambiguous.
But when IBM has its own interpretation of the spec, it's worth to tell it.

That's a feature I like generally in IBM's redbooks, because they clearly make distinction between what's exactly required by the spec, and what's is provided by WSAD as an extra (like for example EJB inheritance or run as delegation not only at bean leval but also method level, etc...)

Regards,
Cyril.

PS: I don't know if god knows truth, in fact, it would be great to know first about its existence...;-)
 
  • Post Reply
  • Bookmark Topic Watch Topic
  • New Topic