db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Joel Rosi-Schwartz <Joel.Rosi-Schwa...@Etish.org>
Subject Re: Integrating Derby and Hibernate
Date Tue, 21 Sep 2004 11:28:46 GMT
On Monday 20 September 2004 17:26, Daniel John Debrunner wrote:
> Joel Rosi-Schwartz wrote:
> > First, though, I think that it might interest the Derby community to know
> > that there is now a DerbyDialect for Hibernate. It was contributed by
> > Simon Johnston and can be found at
> > http://opensource.atlassian.com/projects/hibernate/browse/HB-1224. While
> > this is not yet an official part of the Hibernate release, I have been
> > pounding on it pretty hard for the past two days and I am satisfied that
> > it is good enough for me to start writing code against. It will probably
> > mature when the Hibernate team gets to looking at Derby officially. I
> > believe that they are waiting for an official release from Apache.
> > Personally I think it would be a nice if someone from the team officially
> > contacted Gavin to form a working relationship. IMHO, having a sound
> > integration would be a real win-win proposition.
> >
> > My environment is Cloudscape 10 beta, Hibernate 2.1.6 and Java 1.4.2_04,
> > running on w2k.
> That's great to have the initial Hibernate/Derby integration. Pre-built
> jars are now available on the Derby site so I hope the integration will
> use the Derby driver and JDBC URL names, rather than the temporary names
> from tech preview (beta) IBM Cloudscape release.

In fact this is not an issue. The DerbyDialect simply assists Hibernate in 
creating the correct SQL statements. Hibernate is configurable to so that it 
be used standalone by directly creating connections, in a J2EE web 
environment using the DataSource to retrieve pooled connections, or a full 
blown application server to using JNDI lookup for the database pool. All of 
this is done in the configuration files and is quite straight forward. 
> > The issue is the length limitation of 18 characters that Derby imposes on
> > foreign key constraint names. This is not an issues if one is writing a
> > new schema, but existing schemas and hibernate's auto generation of FK
> > names do cause a problem. Is there anyway to adjust this dynamically (I
> > rather doubt it) or is this essentially a enhancement request?  The IBM
> > Cloudscape Reference Manual indicates that a similar 18 character limit
> > applies to index, function and trigger names. Are these also areas that
> > might cause portability issues?
> This would be a enhancement request, internally Derby can support up to
> 128 characters for identifiers.
> In the mean time can Hibernate address the issue as it does for DB2
> (assuming it does :-), since DB2 has the same limits?

Sorry I do not know the answer to that. I would be best to talk with Gavin at 
the Hibernate projects about the internals.

- joel

View raw message