Hi,

I previously based my hibernate code on another example but found CaveatEmptor more clean. However, in my earlier code the SessionFactory was created in the Filter's init method. In CaveatEmptor I find that you have this initialization as static code in HibernateUtil.

Is this just a matter of choice or are there some benefits/drawbacks with either alternative?

Best regards

Lasse Bergström

The static block works everywhere, not only in a servlet container.

I got it. Thx a lot Christian.

Hi,

I have 2 questions regarding class HibernateUtil:

1) The configuration and sessionFactory objects are created in the static code. How about cleanup of this two objects. When are they null:ed? I am thinking of something similar to a "destroy"-method.

2) There are 2 methods rebuildSessionFactory. When are they used? I can't find any usage of these 2 in the application.

Best regards

Lasse

1) Depends, maybe in a context listener for redeployment? Call sessionFactory.close() when your application "ends".

2) I use it, you don't have to.

Just interested, when do you use rebuildSessionFactory?

When I need to rebuild the SessionFactory after I changed the Configuration.

So you do change the Configuration programmatically?

Why not?

Well I was thinking that normally u add objects manually and the configuration as well. But maybe u are building some kind of interface to edit Hibernate configuration?

Best regards

Lasse

No, but I like to change Configuration settings programmatically if I have to.

Hi,

HibernateUtil.commitTransaction() is called from HibernateFilter. If I want to be able to handle a StaleObjectStateException in my web application and there present a nice message for the user, do I then have to have a HibernateUtil.commitTransaction in DAO-classes as well. Or does the session.saveOrUpdate() method signal a StaleObjectStateException?

The way I do it right now is:

1) In my DAO-class I call HibernateUtil.getSession().saveOrUpdate()
2) In the HibernateFilter the call is made to Hibernate.commitTransaction()

Is this the right way in a web application to be able to get StaleObjectStateException? This way I don't seem to get the optimistic locking strategy to work. My mapping file have this part:



<version name="updateVersion"
column="updateVersion"
type="long"/>


Best regards

Lasse

Hi,

HibernateUtil.commitTransaction() is called from HibernateFilter. If I want to be able to handle a StaleObjectStateException in my web application and there present a nice message for the user, do I then have to have a HibernateUtil.commitTransaction in DAO-classes as well. Or does the session.saveOrUpdate() method signal a StaleObjectStateException?

The way I do it right now is:

1) In my DAO-class I call HibernateUtil.getSession().saveOrUpdate()
2) In the HibernateFilter the call is made to Hibernate.commitTransaction()

Is this the right way in a web application to be able to get StaleObjectStateException? This way I don't seem to get the optimistic locking strategy to work. My mapping file have this part:



<version name="updateVersion"
column="updateVersion"
type="long"/>


Best regards

Lasse

Hi Christian,

From what I understand class HibernateFilter is not actually used in the example since it's not a web application, am I right?

If I am right on that I have some questions.

The call HibernateUtil.commitTransaction() is made from HibernateFilter. Is the purpose of that call to cover for missed call to the same method in the application code?

So, should I put HibernateUtil.commitTransaction after each update, insert, delete etc?

And if I do, do I need the call to HibernateUtil.commitTransaction() in HibernateFilter?

Best regards

Lasse

You can do that in whatever way you like. The purpose is to show you how interceptors can be used for transaction demarcation (and possibly exception handling if you like). Be creative...