Test sakai-2.8.1 Oracle conversion scripts
Description
Environment
Test Plan
depends on
relates to
Activity

Matthew Jones January 3, 2012 at 1:26 PM
I believe everything related to this issue was resolved and 2.8.1 was already tagged awhile ago.

Seth Theriault October 10, 2011 at 9:00 AM
We should recommend that all Oracle users convert the LONGs to CLOBs wholesale (Columbia did this using some old UMich docs).

Matthew Jones October 10, 2011 at 8:55 AM
Thanks Seth, the other error was because I had the wrong Hibernate dialect set, so I think this one can be closed out. It looks like with the 9i dialect it ends up using this? This could be a problem though for Oracle users on 9i who haven't converted all LONG's to CLOBS? It doesn't look like they want to fix it and suggest using a different type? Worked with the 10g dialect though.
http://stackoverflow.com/questions/3719905/why-has-hibernate-switched-to-use-long-over-clob

Seth Theriault October 10, 2011 at 8:42 AM
Fix for SAKAI_MESSAGE_BUNDLE alter statement committed to trunk, r99087. Merged to 2.8.x, r99088, and 2.8.1, r99089.

Matthew Jones August 4, 2011 at 2:38 PM
For some reason, XML in ANNOUNCEMENT_MESSAGE is still "LONG"
alter table ANNOUNCEMENT_MESSAGE modify XML CLOB;
fixes that. This was done to mailarchive and many other tables a while ago ().
And text is not a valid datatype in Oracle for the other statement, it's only trying to change it to null anyway so should be
alter table SAKAI_MESSAGE_BUNDLE modify DEFAULT_VALUE null;
Run Liquibase comparisons against a converted sakai-2.8.0->2.8.1 Oracle 11g database and sakai-2.8.x. Fix issues, clean up scripts.