Re: [Hibernate] Enhancements to the Hibernate Xdoclet subtasks

[Joel Rosi-Schwartz]

I have diff'ed  hibernate-mapping.dtd and hibernate-mapping-1.1.dtd. Many of the changes seem to be because you have tightened up the schema going from #IMPLIED to explicit values in many places. Do any of these require changes to the the hibernate xdt in ...xdocklet.resorces?

The changes that I have found that may need handling are:

1. <joined-subclass> has been added
   
2. the allowed values for <composite-id> has changed
3. <many-to-one length> has been removed
4. <key-many-to-one> has been added
5. <map generated-key> has been removed
6. <array generated-key> has been removed
7. <many-to-many length> has been removed
8. <key type> has been removed
10. <generated-key> has been added
11. <index-many-to-many> has been added
12. <cache> has been removed
    

Is this list complete? Which of these changes require addition, removal or change to the xcdoclet subtasks?

I can not find any reference to composite-id in hibernate xdoclet, so I am assuming that it something that needs support added rather than modified. Is this accurate?

I am going to move on and look at how the hibernate dtd correlates to the xdoclet dtd, but some guidance and insight will save me some time and effort. In the meantime I am finding it valuable to gain an understanding of the hibernate schema, I am certain that will have a payback later in my projects. One moan however, I wish you had used Xschema I am really rusty at dtd. (This is not intended to start a meaningless thread on the pro and cons of using dtd vs. xschema, there are volumes filled with that one already ;-)

 - joel

Gavin King wrote:

Firstly and most importantly, the XDoclet module needs to be updated to
support the new DTD (hibernate-mapping-1.1.dtd). The biggest and most
important change in this is support for the <joined-subclass> mapping style.
However there are also a couple of smaller changes (eg.
<composite-id><property> was changed to <composite-id><key-property>,
<composite-id><key-many-to-one> ). I don't expect it would be very difficult
to extrapolate these changes from the current functionality.
<joined-subclass> is substantially similar to <subclass>, with the exception
that it has the required <key> subelement.

Gavin