|
||||||||
|
This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira |
||||||||
|
||||||||
|
This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira |
||||||||
I think you're probably right about checking the delayed status in BOTH the entry and exit. The exit case would simply clear(). However, if the timeout somehow happens in the entry case, what should we do? clear() and prevent the rest of the method from occurring?
We could rename checkTransactionSynchStatus and have it also call a delayedAfterCompletion method in SynchronizationCallbackCoordinatorImpl. If the delayed status flag had been set, call afterCompletion. SessionImpl is already registered as a TransactionObserver, so it would receive the completion and then clear itself. Although, wouldn't we still need to do something to stop the processing of the original Session method call?