Re: Completeness test with different relationship configurations

[Craig Russell]

Hi Michael, Mea culpa. I did not realize that when I changed the test data of the company model to remove the relationships, I also removed the entire graph except for the company1 model. That single instance was persisted, and that single instance was compared. I had meant that all of the objects in the graph would have been instantiated and made persistent and compared. So your concern with the 1-1 relationships below is just part of a bigger issue. I think the solution is to identify the "islands" of non-connected object graphs and to check all of the islands in the comparison. The list of "islands" would need to be specified as a property, as a file, or hard coded into subclasses of CompletenessTest.  Craig On Jul 7, 2005, at 8:42 AM, Michael Watzek wrote: Hi Craig, currently, I'm preparing the completeness test to run with different configurations as we decided in last t-conference. I started with adding input XML files for the company model for different relationship types: - companyEmbedded.xml: This input file adds embedded objects to input file companyNoRelationships.xml. - company1-1Relationships.xml This input file adds 1-1 relationships to input file companyNoRelationships.xml. - company1-MRelationships.xml This input file adds 1-M relationships to input file companyNoRelationships.xml. - companyM-MRelationships.xml: This input file adds M-M relationships to input file companyNoRelationships.xml. My understanding is that we add a new configuration for each of these input files, and that we run the completeness test with each of those configurations. Currently, the completeness test reads one root object (company1) from an input file and makes that persistent. Afterwards, it starts a new transaction and reads that root object from the database. Finally, it compares the two objects. I think, we have to adapt the test case in order to make 1-1 relationships work: Due to the fact that company does not have a 1-1 relationship, the XML reader only creates an isolated company object in this case. I see two options to solve this: 1) We pass the name of a root object to the test case and add the name to the configurations. Different configurations would create different root objects which are appropriate for their needs. 2) We change the completeness test in order to create one root object of each persistent model class. What do you think? Regards, Michael --  ------------------------------------------------------------------- Michael Watzek                  [EMAIL PROTECTED] Engineering GmbH mailto:[EMAIL PROTECTED]        Buelowstr. 66 Tel.:  ++49/30/235 520 36       10783 Berlin - Germany Fax.:  ++49/30/217 520 12       http://www.spree.de/ ------------------------------------------------------------------- Craig Russell Architect, Sun Java Enterprise System http://java.sun.com/products/jdo 408 276-5638 mailto:[EMAIL PROTECTED] P.S. A good JDO? O, Gasp!

smime.p7s
Description: S/MIME cryptographic signature