Dear All, I have a Turbine application that has a schema that described all the metadata for schemas. So a user can describe a database they have and we have existing data files that we map values from into the database they have described using the screens in the application. Ideally what I would like to do is: 1. allow the user to fill in the data about their database and table structures and click a button. 2. which kicks off the application to run a service which takes the metadata and generates a schema in XML 3. then run ant init for that schema (call org.apache.tools.ant.Main() would seem to work) to generate the database (optional, it may exist) and generate the OR Peer classes. 4. start using the OR Peer classes (This is tricky I think as the ClassLoader won't be able to see these classes, noy sure if I can sub class ClassLoader to be able dynamically load the new package) with the Database connection and driver details also entered in step 1 above.
The main point is that if I can achieve this without having to restart the application it would prevent any ScheduledJobs it has running (it runs a lot) being terminated. The main problems I can see is how to name a package that keeps different databases distinct and yet be able to address these classes from the existing code. My first thought is to use the Bridge pattern to isolate any concrete OR Peer classes from the abstraction I hold in the metadata. Ideas on approach or solutions for this? Thanks David -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
