On Sep 15, 2004, at 5:18 AM, Jacek Laskowski wrote:
[EMAIL PROTECTED] wrote:
I am trying to get a better understanding of database support in Geronimo (and its components) and have a number of questions:
* The page http://openejb.codehaus.org/about.html indicates that OpenEJB provides EJB1.1 compatibility, but on the page http://openejb.codehaus.org/cvs.html it indicates that the version of OpenEJB for Geronimo has EJB 2.x support.
This is because there're two versions of OpenEJB, the one being integrated with Geronimo - it's OpenEJB2, and the one that has some nice features like Tomcat integration with (only) EJB 1.1 + Local interfaces support - it's OpenEJB1.
Only OpenEJB2 will support EJB 2.x and more.
To clarify, OpenEJB 1 was written to be compatible with the EJB 1.1 specification, although it is not certified (much like JBoss was prior to version 4 which has claimed certification).
OpenEJB 2 has been refactored (and added to) to be EJB 2.0 compliant and integrated into Geronimo.
* Whilst trying to get a picture of the current state of OpenEJB, I noticed that there are JIRA issues for OpenEJB on both the Geronimo and the http://openejb.codehaus.org/ site.
When is each site used for raising issues? (e.g. is the Geronimo JIRA database only used to record OpenEJB problems that impact integration of OpenEJB into Geronimo or passing TCK tests)?
As a OpenEJB committer I can say I treat these two JIRA repos as dictinct: one for OpenEJB2 (it's the JIRA installation hosted at Apache.org) and the another for OpenEJB1 hosted at CodeHaus.org.
So, issues shown while working with Geronimo (and OpenEJB2) are going to JIRA at Apache.org, whereas with OpenEJB1 (which is not expected to work with Geronimo) to JIRA at CodeHaus.org.
Yep. We are entering the OpenEJB issues into the Geronimo JIRA so we can keep track of the overall progress of Geronimo from a single location.
* Geronimo uses Castor version 0.9.3.9, which is dated December 11, 2001. Is there a reason why more recent versions are not used? ( http://www.castor.org/changelog.html )
I don't know about any. I also think you're talking about OpenEJB2 (or even OpenEJB1), weren't you? If so, that's been overlooked. It seems that noone noticed Castor upgrade and it didn't go to OpenEJB.
Geronimo itself doesn't currently use Castor, but some parts of OpenEJB does. When we create the Geronimo distribution we need to include castor for OpenEJB.
* If I want to use Geronimo with a proprietary database that has a JDBC driver, is Castor the only library that may need modification to work with the proprietary database? What about TranQL?
Well, I'm confused. I don't think Castor is used in Geronimo, but OpenEJB. I might have missed something :( As far as TranQL goes, I thought, it's used for CMPs. JDBC drivers are handled by openejb-jca wrapper, but wouldn't cound on the information.
Jacek was close. In Geronimo, JDBC connections are managed using the J2EE Connector APIs. We currently use the tranql-connector JDBC to JCA wrapper in Geronimo. We used to use the openejb-jca wrapper, but dropped it since it is only JCA 1.0 and the Tranql one is JCA 1.5, and openejb-jca one is copyright Intalio (long story there).
-dain
