Actually looks like we can reproduce this problem. EJBQLQuery does
query compilation on the client, and that accesses DbRelationships. So
this is not related to naming, but rather to the use of joins.
A fix can be rather invasive, so how does everybody feel about making
EJBQL a server-only feature in 3.0 and doing necessary refactoring in
3.1?
Andrus
On Jan 26, 2010, at 5:03 PM, Andrus Adamchik wrote:
>
> On Jan 26, 2010, at 4:50 PM, Lachlan Deck wrote:
>
>> Strictly speaking, yes. But you could also say that one is a
>> derivative of the other. e.g., upper-casing camel casing and joined
>> via underscore which is a common db mapping (which perhaps cayenne
>> allows for, especially when reverse engineering).
>
> No, Cayenne doesn't make any such adhoc conversions. So MT_TABLE and
> MtTable1 are different names, and the framework will not attempt a
> conversion.
>
> So do you have a stack trace of the error BTW?
>
> Andrus
>
This archive was generated by hypermail 2.0.0 : Tue Jan 26 2010 - 10:26:15 EST