Emmanuel Cecchet <[EMAIL PROTECTED]> writes: >> Where should we start solving this?
> I think that the right place for solving this would be the > SQLExceptionFactory class where we systematically chain the original > exception where we should probably just take the message. That should do it. > Can you also file a JIRA entry in Sequoia for this? Unfortunately that would harm the controller->driver communication which has been designed to take care of this and is totally immune to such issues. Dropping exception-related information always makes debugging and diagnosing infinitely harder. I'd prefer fixes along those lines: - simplify class-loading. https://forge.continuent.org/jira/browse/SEQUOIA-751 - specifically catching this serialization problem (blindly catching all exception types at once is never a good idea; exception types do not exist just for aesthetic reasons) and then implement a "graceful degradation" (!= hanging) for this specific problem. _______________________________________________ Sequoia mailing list Sequoia@lists.forge.continuent.org https://forge.continuent.org/mailman/listinfo/sequoia
