I work with Naga and have been following this thread.  

It sounds like the SQL generation of foreign key constraints in schema
xml files is wonky.  

Specifically: when the foreign-key's attribute foreignTable refers to a
table alias, it's not finding the real table name and substituting it
into the "ALTER TABLE" sql.

Do I have a correct understanding of the problem?

+jeff

-----Original Message-----
From: Blake Day [mailto:[EMAIL PROTECTED]]
Sent: Thursday, May 09, 2002 9:16 AM
To: Turbine Users List
Subject: RE: ant init error messages


Actually, I was manually building the databases at first until I tried
changing the name of the table.  That worked, but it certainly was
annoying.
I was running ant init, but pausing right after the SQL was generated so
I
could go make the changes to the file.  I then unpaused ant init for the
SQL
insertion to take place.

Michael Blake Day
Artistry Studios - e-commerce design, implementation and hosting
email: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>


-----Original Message-----
From: Colm McCartan [mailto:[EMAIL PROTECTED]]
Sent: Thursday, May 09, 2002 4:53 AM
To: Turbine Users List
Subject: Re: ant init error messages


At 09:46 09/05/2002 +1000, you wrote:
><disclaimer>I haven't read this thread thoroughly.</disclaimer>
>
>The extend-user-howto was put together using MySQL which ignores the
>referential integrity constraints that are referred to in the original
post.
>It is indeed possible that the documented solution will require
modification
>for an RDBMS that supports RI.
>
>Have others using PostgreSQL or Oracle successfully extended
TurbineUser in
>the manner described in the HOWTO?
>
>Scott

Yes, we managed this with postgres. I posted earlier in the thread
saying
that this alias name is invalid with postgres because of RI - we got
around
this by referring the foreign key directly to TURBINE_USER. I thought
Blake's solution of using the table's java-name attr to point to the
appropriate class sounded like a good idea - we ended up generating the
code and accepting the eerrors and then manually building the database.

Re: the primary key violations that Oracle is reporting: are these
tables
being correctly dropped/depopulated during the build process?

colm


--
To unsubscribe, e-mail:
<mailto:[EMAIL PROTECTED]>
For additional commands, e-mail:
<mailto:[EMAIL PROTECTED]>


--
To unsubscribe, e-mail:
<mailto:[EMAIL PROTECTED]>
For additional commands, e-mail:
<mailto:[EMAIL PROTECTED]>


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to