On 30/12/2013 21:44, Oliver Wulff wrote:
Hi there I'm looking into adding a persistence layer to the Apache Fediz Identity Provider and used the persistence layer in Syncope as a good reference. Nevertheless, I got a few questions...
Hey, looks cool :-)
1) JPA creates the tables when the first entity is added - usually triggered by a REST call. They won't be created when the spring application context is loaded. The Syncope ContentLoader does only add rows to existing tables and creates indexes and views, but no tables. How did you workaround this?
OpenJPA is configured in Syncope to create tables when Spring context is loaded (you can experiment this by providing an empty content.xml or by disabling the ContentLoader).
2) What was the reason for you to create an orm.xml instead of using annotations only?
The general policy was to use orm.xml for all configurable parameters that would have been instead stuck as annotation parameters in the source code. For example: putting table generators in orm.xml lets you easily configure the initial value to pick without changing the source code. This is very useful in Syncope since it is meant to be used via WAR overlays.
3) You introduced your own semantic of a file to import default data (content.xml). I assume you did not use an SQL script to be independent of the database?
Correct.
But why did you use the JDBC template instead of the JPA layer?
Hum, that's a nice question, probably related to the evolution that ContentLoader had from original (pre-ASF) DBUnit. I guess that one could rewrite ContentLoader, ContentUpgrader and ContentExporter to work with EntityManager and thus remove spring-jdbc from main dependencies (it will still be used by integration tests code).
Regards. -- Francesco Chicchiriccò Tirasa - Open Source Excellence http://www.tirasa.net/ ASF Member, Apache Syncope PMC chair, Apache Cocoon PMC Member http://people.apache.org/~ilgrosso/
