[ http://opensource.atlassian.com/projects/hibernate/browse/HHH-6726?page=c... ] Gail Badner commented on HHH-6726: ---------------------------------- After thinking about this more, I think we should keep the Oracle dialects as is. Here are my reasons: - Hibernate defines the text/image types consistent with how Oracle maps LONGVARCHAR/LONGVARBINARY; - the bug lies with Oracle; - Oracle has effectively deprecated valid SQL data types (LONGVARCHAR and LONGVARBINARY); - the migration from "text" to "materialized_clob" has already been documented since 3.5; "image" is a new type as of 3.6; - anyone actually using text/image with LONG/LONGRAW columns may run into new migration issues in Hibernate 4.0.0; more testing would need to be done to determine the true impact; - users can easily create a new dialect making the changes I mentioned earlier; - this avoids having to make a confusing recommendation (i.e., using Oracle11gDialect for accessing an Oracle 10g DB). You can create a custom dialect to make these changes by extending Oracle10gDialect and adding the following to the constructor: registerColumnType( Types.VARBINARY, 2000, "raw($l)" ); registerColumnType( Types.VARBINARY, "blob" ); registerColumnType( Types.LONGVARCHAR, "clob" ); registerColumnType( Types.LONGVARBINARY, "blob" ); registerColumnType( Types.VARCHAR, 4000, "varchar2($l char)" ); registerColumnType( Types.VARCHAR, "clob" ); -- This message is automatically generated by JIRA. For more information on JIRA, see: http://www.atlassian.com/software/jira