[
https://issues.apache.org/jira/browse/JDO-521?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12525981
]
Andy Jefferson commented on JDO-521:
------------------------------------
The "negative" 1-1 relation tests raise a question. Let's take the first one.
You have an already persistent Employee, and an already persistent
MedicalInsurance. You then persist a new Employee. Ok, no problem with that.
You then call
emp1.setMedicalInsurance(medIns2);
The spec says "In particular, making a change to one side of the relationship
does not imply any runtime behavior by the JDO implementation to change the
other side of the relationship in memory prior to flush"
but that doesn't by my reading prohibit updating the in memory model, and so
for any implementation that does update in-memory on each change these
"negative" tests will fail since they don't allow for it. Or is it that there
is some requirement to prohibit this ? An implementation *can* flush changes to
the datastore without user-called flush() or commit() (unless using optimistic
txns).
PS. Any chance the test "conf" files can be called "managedRelations.conf" or
something ? "relationshipsNoRelationships.conf" leaves me baffled ... since if
it is for relationships and has no relationships?? surely it would have to have
some relationships to test managed relationships ? Maybe I'm missing some
secret naming schema
> Create tests for relationship mapping (spec section 15.3)
> ---------------------------------------------------------
>
> Key: JDO-521
> URL: https://issues.apache.org/jira/browse/JDO-521
> Project: JDO
> Issue Type: Test
> Components: tck2
> Affects Versions: JDO 2 maintenance release 1
> Reporter: Michelle Caisse
> Assignee: Michelle Caisse
> Attachments: JDO-521.patch
>
>
> We need a set of tests to test that an implementation syncs both sides of a
> relationship on flush().
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
