You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@turbine.apache.org by Colm McCartan <co...@owl.co.uk> on 2002/01/07 12:25:30 UTC
Postgres issues (Was:Re: adding new types to Torque..mysql)
John McNally wrote:
> Sorry, I missed the response to my question, but you should specify the
> fk as a composite, not as two distinct fk's.
>
>
>>> <foreign-key foreignTable="DOCUMENT_VERSION">
>>> <reference local="DOCUMENT_ID" foreign="DOCUMENT_ID"/>
>>> </foreign-key>
>>> <foreign-key foreignTable="DOCUMENT_VERSION">
>>> <reference local="VERSION_ID" foreign="VERSION_ID"/>
>>> </foreign-key>
>>>
>
> Should be:
>
>
>>> <foreign-key foreignTable="DOCUMENT_VERSION">
>>> <reference local="DOCUMENT_ID" foreign="DOCUMENT_ID"/>
>>> <reference local="VERSION_ID" foreign="VERSION_ID"/>
>>> </foreign-key>
>>>
>
> john mcnally
>
> Daniel Rall wrote:
>
>>Hi Colm. Was there ever a resolution to this?
OK. Thanks for the clarification John - I shall try this at my next
rebuild. BTW, what are your thoughts on the bad SQL generated for the
user table alias as discussed in:
http://www.mail-archive.com/turbine-user@jakarta.apache.org/msg05399.html
Here, foreign key references are made to the table alias name, instead
of to the underlying turbine_user table...
Also, I am tracing some problems with native id generation. The
DBPostgres adapter class appears to have invalid code in its
getIdSqlForAutoIncrement method. It seems to try to issue SQL that looks
like this:
select currval(seq_name);
instead of
select last_value from seq_name;
Has anyone come across this?
All in all, I'm tempted to go over to idBroker to bypass this problem.
Any thoughts welcome.
While I'm on, thanks for your comments Scott (Eade) on how to approach
patching Torque - at the moment this may demand more time than I can
afford but cheers anyway. I would welcome any details about the torque
test framework you mention. Also, you mentioned a new 'hack' for
Object.vm - I'd certainly like to see it included in the howto.
Cheers,
colm
--
To unsubscribe, e-mail: <ma...@jakarta.apache.org>
For additional commands, e-mail: <ma...@jakarta.apache.org>