The code that generates the DBUnit XML file is really robust and is a great time-saver! Thank you so much for writing it. However, we noticed that the generated file is missing the many-to-many join tables. This was pretty simple to fix though.
In dbhelper.ext:
// get unique list of join tables
Set[String] getJoinTableNames(Collection[DomainObject] domainObjects) :
domainObjects.collect( d | d.getJoinTableNames()).flatten().toSet();
// get join tables for this domain object
Set[String] getJoinTableNames(DomainObject domainObject) :
domainObject.references.select(r | isManyToMany(r)
&& r.to.hasOwnDatabaseRepresentation())
.getManyToManyJoinTableName();
In the template:
«DEFINE dbunitTestData FOR Service»
«FILE "dbunit/" + name + "Test.xml" TO_RESOURCES_TEST»
<?xml version='1.0' encoding='UTF-8'?>
<dataset>
«LET module.application.getDomainObjectsInCreateOrder(true) AS domainObjects -»
«FOREACH domainObjects AS domainObject »
<«domainObject.getDatabaseName()» />
«ENDFOREACH »
«FOREACH domainObjects.getJoinTableNames() AS joinTableName »
<«joinTableName» />
«ENDFOREACH »
«ENDLET»
</dataset>
«ENDFILE »
«ENDDEFINE »
The code that I tested with is slightly different due to my customizations, but that should be what is needed for Sculptor trunk code. Hope that helps.
The code that generates the DBUnit XML file is really robust and is a great time-saver! Thank you so much for writing it. However, we noticed that the generated file is missing the many-to-many join tables. This was pretty simple to fix though.
In dbhelper.ext:
// get unique list of join tables
Set[String] getJoinTableNames(Collection[DomainObject] domainObjects) :
domainObjects.collect( d | d.getJoinTableNames()).flatten().toSet();
// get join tables for this domain object
Set[String] getJoinTableNames(DomainObject domainObject) :
domainObject.references.select(r | isManyToMany(r)
&& r.to.hasOwnDatabaseRepresentation())
.getManyToManyJoinTableName();
In the template:
«DEFINE dbunitTestData FOR Service»
«FILE "dbunit/" + name + "Test.xml" TO_RESOURCES_TEST»
<?xml version='1.0' encoding='UTF-8'?>
<dataset>
«LET module.application.getDomainObjectsInCreateOrder(true) AS domainObjects -»
«FOREACH domainObjects AS domainObject »
<«domainObject.getDatabaseName()» />
«ENDFOREACH »
«FOREACH domainObjects.getJoinTableNames() AS joinTableName »
<«joinTableName» />
«ENDFOREACH »
«ENDLET»
</dataset>
«ENDFILE »
«ENDDEFINE »
The code that I tested with is slightly different due to my customizations, but that should be what is needed for Sculptor trunk code. Hope that helps.
Sure! Thanks for pointing out CSC-382. I did not see that. We already have code that drops all constraints, runs the delete-all, and then reapplies the constraints, but your solution looks more database agnostic. We may need to switch to using your solution.
Thanks!
--Polly
Patrik Nordwall wrote:
Thanks a lot for your contribution Polly. I will include it in trunk.
