Hello,

I'm trying to better understand internals of the reverse engineering strategies. So I've made a class (extending DelegatingReverseEngineeringStrategy) that override the DefaultReverseEngineeringStrategy.
In this class I've defined a method foreignKeyToEntityName that is the one-to-one copy of the same method contained in the class DefaultReverseEngineeringStrategy.

public String foreignKeyToEntityName(String keyname, TableIdentifier fromTable, List fromColumnNames, TableIdentifier referencedTable, List referencedColumnNamesString, boolean uniqueReference) {
String propertyName = Introspector.decapitalize(StringHelper.unqualify(tableToClassName(referencedTable)));

if(!uniqueReference)
if(fromColumnNames != null && fromColumnNames.size() == 1)
{
String columnName = ((Column)fromColumnNames.get(0)).getName();
propertyName = propertyName + "By" + toUpperCamelCase(columnName);
} else
{
propertyName = propertyName + "By" + toUpperCamelCase(keyname);
}
return propertyName;
}

I was surprised to see that results that I've obtained are not the same....

For example, I've a table called "CONTATTI" containing two one-to-many relationships having the same target table.
If I don't specify a class for the reversestrategy attribute I obtain this result:

<many-to-one name="tipocontatto" class="it.csi.ccrp_bo.hibernate_stat.Tipocontatto" fetch="select">
<column name="IDTIPOCONTATTO" precision="22" scale="0" not-null="true" />
</many-to-one>
<many-to-one name="tipocontatto_1" class="it.csi.ccrp_bo.hibernate_stat.Tipocontatto" fetch="select">
<column name="IDTIPOCONTATTO_OUT" precision="22" scale="0" />
</many-to-one>

These results are the same that I obtain with Hibernatetools called directly from Eclipse.

Using my own reversestrategy class I obtained:

<many-to-one name="tipocontattoByIdtipocontatto" class="it.csi.ccrp_bo.hibernate_stat.Tipocontatto" fetch="select">
<column name="IDTIPOCONTATTO" precision="22" scale="0" not-null="true" />
</many-to-one>
<many-to-one name="tipocontattoByIdtipocontattoOut" class="it.csi.ccrp_bo.hibernate_stat.Tipocontatto" fetch="select">
<column name="IDTIPOCONTATTO_OUT" precision="22" scale="0" />
</many-to-one>

Now the questions are:
Why in one case I obtained 'tipocontatto' and 'tipocontatto_1' and in the other I obtained those different (and I think more correct) results ?
What is the class used If I don't specify a reversestrategy one ? I thought that is was DefaultReverseEngineeringStrategy but obviously I were in mistake.

Thanks in advance for you support.
Janniz

users have reported this before but noone have never given me a testcase that reproduces this on hibernate tools beta's.

So please do by putting it in jira.

That sounds very similar to my issue HBX-467 (which include everything
neeed to reproduce the problem).

George

No - your issue is about how the naming strategy actually works, and that case actually shows that the codegenerator does not append _1,_2 ...

Ok, I'll post it on Jira on my lunch time this afternoon.
What do you need to make a testcase ? (Hibernate hbm.xml ? ER diagram of the db or what else ?)

Thank you
Janniz

i need a build.xml and the source of your custom strategy so i can reproduce the case.

In short the actual settings/files needed to reproduce it.