Yeah, the example is a little odd, but combined with runtime relationship
created in CAY-1009 for base types, the situation does come up. A fix of
CAY-1009 may mitigate the need for this. In which case, we could either
explicitly disallow it or come up with a better alternative.
Off-hand, I'm thinking a means of mapped queries that take the qualifier
into account and cast results to the appropriate type. I suppose this would
have more performance implications than just maintaining the list locally,
but it should take care of any data integrity issues without getting too
ugly.
On 3/15/08 2:13 PM, "Andrus Adamchik" <andru..bjectstyle.org> wrote:
>
> On Mar 15, 2008, at 7:31 PM, kmenar..pache.org wrote:
>
>> <obj-relationship name="addresses" source="Employee"
>> target="Address" db-relationship-path="employeeAddresses"/>
>> + <obj-relationship name="homeAddresses" source="Employee"
>> target="HomeAddress" db-relationship-path="employeeAddresses"/>
>
> Hmm.... I know for sure I never had such mapping (which explains why
> the bug is there to begin with). Still formally the mapping is valid,
> so I guess we'd have to handle it.
>
> Andrus
>
This archive was generated by hypermail 2.0.0 : Sat Mar 15 2008 - 16:13:35 EDT