Thanks for clarifying.
Perhaps I can clarify the difference with two pseudo-ish code examples:
(1)
@Stateless
public class RepositoryEJB {
@Resource private Repository repository;
public String getMimeType(String path) {
Session session = repository.login( .. );
String mimeType =
(String)session.getRootNode().getNode(path).getProperties().get( .. );
session.logout();
return mimeType;
}
public Long getSize(String path) {
Session session = repository.login( .. );
Long size =
(Long)session.getRootNode().getNode(path).getProperties().get( .. );
session.logout();
return size;
}
}
(2)
@Stateless
public class RepositoryEJB {
@Resource private Repository repository;
private Session session;
@PostConstruct
public void postConstruct() {
Session session = repository.login( .. );
}
@PreDestroy
public void preDestroy() {
session.logout();
}
public String getMimeType(String path) {
return (String)session.getRootNode().getNode(path).getProperties().get(
.. );
}
public Long getSize(String path) {
return (Long)session.getRootNode().getNode(path).getProperties().get( ..
);
}
}
Which one is better?
Thank you,
Jaco
On Wed, Oct 14, 2009 at 11:04 PM, Guo Du <[email protected]> wrote:
> On Wed, Oct 14, 2009 at 11:11 AM, Jaco Prinsloo <[email protected]>
> wrote:
> > 1) Every request (method) opens a session, performs the work and then
> closes
> > the session again.
> In general, we have a session open to process the request, works like
> jdbc data source. It doesn't matter what environment you are using,
> servlet container, ejb or standalone application.
>
> > 2) There is one session per SLSB which is opened in PostConstruct and
> closed
> > in PreDestroy.
> This is how you actual implement the session management in EJB container.
>
> > Are there any advantages/disadvantages to these two approaches?
>
> I couldn't see the difference in your approaches.
>
> -Guo
>
