Hibernate version:
3.1.3
Hibernate tools version:
3.2.2 beta 1

This is not an error rather I am asking for suggestions on how to work around manual changes.

I have the following setup which is running in production right now.
Around 60 tables with diverse contents.
From these tables are generated hbm files and POJO's.
But I have a number of issues which I have not been able to configure when generating hbm files. The issues are concerned with having to manually change the hbm files after database generation.

1. I do not want my hbm files to contain the name of the catalog and schema. The reasone for this is that I have a number of different databases both in production and test which have different names. Is there a way to specify that we do not want these attributes in the hbm files?

2. I have been able to set the generated class attribute for all tables, using a CustomRevengStrategy and meta attributes. This does not work for classes with composite id's however . Here I have to specify manually the name of the composite id's base class in the meta attribute. Is there anyway to do this automatically?

3. I have a number of property fields which are based on formulas in the hbm files. These would be nice to write in the hibernate reverse engineering file like in some way like; "....
<table name="XXX>
<column name="xxx" type="xxx" formula="xxxx"/>
</table>
But unfortunately this is not supported... Is there some other way to do this?

4. On some of the tables is a <timestamp> property.
These are generated from hibernate as a <property name="xxx" type="timestamp">... instead of <timestamp name="xxx" coloumn="Xxx" />. Is there some way to control this?

set default_schema and default_catalog and if the catalog/schema matches it won't end up in the mapping files.



Please report with instructions on how to reproduce in jira - sounds like a bug.



hmm..ever considered writing mapping files ? :)

reveng.xml is mostly about the relational info, but I guess we could add this in...just not sure sure how yet. Put it in jira.

p.s. you can always add it to the templates manually.,,


you can write a custom revengstrategy that identifies xxx as the optimistic lock column.

Where do you set this? Which file?


This is not a bug, it works fine, but i just cant create the second meta attribute in the below hbm file.
My hbm files look like this;
<class name="XXX1" table="XXX1">
<meta attribute="generated-class">XXX1Base</meta>
<composite-id name="XXX2" class="XXX2">
<meta attribute="generated-class">XXX2Base</meta>
.....

The first meta attibute is constructed by using a custom reveng strategy class where the method tableToMetaAttributes() is overwritten.
But can i do the same with the composite id meta attributes?



The thing is we really want the mapping files and pojos to be genererated automatically by reverse enginering and that is why i would like this ;-).

How would you do it with a template? I havent used these before...

OK so I just override the method "getOptimisticLockColumnName" and return i.e. modifiedDate as a string if that is the name of our timestamp coloumn?

Thanks for you reply btw...
/Anders

This is not a bug, it works fine, but i just cant create the second meta attribute in the below hbm file.
My hbm files look like this;
<class name="XXX1" table="XXX1">
<meta attribute="generated-class">XXX1Base</meta>
<composite-id name="XXX2" class="XXX2">
<meta attribute="generated-class">XXX2Base</meta>
.....

The first meta attibute is constructed by using a custom reveng strategy class where the method tableToMetaAttributes() is overwritten.
But can i do the same with the composite id meta attributes?


ah...we might not have that covered...not sure where to put that info in reveng.xml either...anyway report it in jira - we might need a compoentToMetaAttributes or something.



The thing is we really want the mapping files and pojos to be genererated automatically by reverse enginering and that is why i would like this ;-).
[/quote]

well...come contribute patches for the missing usecases then ;)



yes, that is one way.

As a result of this thread i have opened two Jira issues:

http://opensource.atlassian.com/project ... e/HBX-1081
http://opensource.atlassian.com/project ... e/HBX-1082