We're going to generate a BaseObject class and create the database using the Torque build scripts using a custom project-schema.xml file. Additionally, we're going to alter the Velocity scripts to add in ojb-xdoclet tags to these classes. We're going to have an extended class also created (never overwritten, just created if it doesn't exist). This class is where we'll put out 1:n / n:m collections (manually). These will have ojb-xdoclet tags as well... The final step is to then use the ojb-xdoclet parser to create the repository_user.xml file for OJB.
Does this sound like a viable solution? Am I missing something?
Basically, our goal is to have as much generated from a config file as possible... We'd like to not have to create each class that we know we're going to need... This seemed like a good alternative to the pure ojb-doclet approach with required a decent amount of hand-coding...
Any suggestions would be greatly appreciated.
On Thursday, June 12, 2003, at 01:34 AM, Danilo Tommasina wrote:
hi,
I've looked at using the OJB-XDoclet project to do the same type of
thing, but it seems to have some bugs as well. Plus, I prefer to do my
modeling from the DB, and just extend those base classes.
yep, we also have to work with reverse engineering, we are using torque because it gives us much more control on the repository/code generation.
we also need to do some automated changes in the repository (after generating it trough torque) before generating the object model.
The reverse db tool doesn't give us the possibility to do that and to integrate these steps in ant scripts.
As far as working with 1:n / n:m, I think that we're going to extend
the Torque DTD/schema and add a new possible data type "VECTOR". Then
using the Velocity templates build specialized sections when using this
data type. We've already added some base functionality through the
Velocity templates, but haven't tried this yet.
Let me know if this approach sounds good to you...
yep, i was also thinking to do something similar, however I still hadn't the time to check it deeply (just writing some docu for the project, boooooring)
It is probably useful to get a look how the reversedb tool generates the shema for 1:n/n:m relationships by looking in its source code (yeah, I love Open Source)
I'm probably going into this in the second half of next week, so if you have some solutions, it would be great if you could post them. Thanks
bye danilo
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
-=-= -=-= -=-= -=-= -=-= -=-= -=-= -=-= -=-= -=-= -=-=- Marcus Breese [EMAIL PROTECTED] IU School of Medicine [EMAIL PROTECTED] Dept. of Biochemistry and Molecular Biology Center for Medical Genomics / Grow Lab
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
