At 01:25 PM 7/28/99 , you wrote:
>Should not most find* methods be TX_SUPPORTED? That is, you don't need a
>transaction to query, but the query could be part of a transaction.
>
>This would seem to apply to all non-state modifying methods, such as
>getters.
That makes sense.
>All state modifying methods, on the other hand, should require some form of
>transaction.
>
>Is this correct?
>
>Weblogic does not seem to support the TX_SUPPORTED value. Why?
This is actually a problem with the base classes Sun ships
(javax.ejb.deployment.ControlDescriptor). Note the missing case in the
method setTransactionAttribute. Every vendor should have this problem.
>Is the next best thing using TX_REQUIRED but with the lowest isolation
>(TRANSACTION_READ_UNCOMMITTED)?
Depends on the situation, but this is reasonable.
>Am I totally confused?
No more than most.
Greg
>===========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff EJB-INTEREST". For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".
----------------------------------------------
Greg Desmarais
Practice Manager, WebLogic
BEA Systems
(ph) 617-742-8045
(fax) 617-720-3598
(email) mailto:[EMAIL PROTECTED]
(www) http://www.beasys.com
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
