Hi,

my project involves taking an EJB 2.x app and migrating(re-writting) it to EJB 3.0 using hibernate, where the older app used CMP & xdoclet for some code generation.

If somebody could offer steps on how I could go about this, would be great.

hi!

here are some steps which you want to try:

1st: read the ejb3.0 and hibernate persistence specifications which can be found at the hibernate.org

2nd step: reverse engineer database tables. use the hibernate tools plugin for eclipse which supports ejb3 annotations. if you have a proper designed database-schema then the plugin recognizes the primary keys, foreign keys and their related associations. there are some ant tasks which do that for you, too. within this step you need to configure the connection to the database within a "hibernate-config-file". a reverse-engineering file could be useful. docu for all that features can be found under hibernate.org as mentioned.

3rd step: migrate the rest of your application using the right annotations.

xdoclet is not required any long when using ejb3.0 and hibernate.

hope that helps in some way.

best regards
patrik

ps.: a sample sourcecode can be found at http://caveatemptor.hibernate.org

Thank you for your reply, I am still in the dark about how xdoclet would not be required though, will study the resources that you have cited. Would the app still work if I left the original Xdoclet annotations in there while adjusting it for EJB 3.0 and hibernate?

Thank you

i don't know exactly, i haven't try something like this before...

i guess they're ignored because they're within comments...

It must be easy to migrate if application is designed using "value object" and "session facade" stuff. Just drop entity beans, annotate "value object" and write queries using EntityManager API in "facade" implementation. If you access entity beans directly then it must be a problem to migrate, probably you will need to redesign application.