"Louth, William (Exchange)" wrote:
>
> Do people think that this is breaking the container contract, just working
> around it, enhancing it? I need more time to consider what the implications
> for these features are. I hope eventually that we can come up with a
> standard way (specification ?.0) to address the real world needs that led to
> these features being developed.
The container should definitely call ejbLoad and ejbStore at the
transaction boundaries. That is the only way for the entity bean to
engage in any persistent-related behavior.
That does not mean that CMP container has to load anything (it can use
the cache) or store anything (no changes made). But that's a decision
the CMP container can take regarding the fields it manages and only
these fields.
It's possible that my bean will be doing something else in ejbLoad and
ejbStore that is not affected by caching or is-modified, and any missing
ejbLoad / ejbStore calls will break my bean.
This is a consistent container - component contract that remains
consistent regardless of any particular optimizations that CMP
containers (but not BMP containers, and not all CMP containers in the
same manner) may do.
arkin
--
----------------------------------------------------------------------
Assaf Arkin www.exoffice.com
CTO, Exoffice Technologies, Inc. www.exolab.org
===========================================================================
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".
