diff options
| author | Martin Ritchie <ritchiem@apache.org> | 2009-09-28 15:33:18 +0000 |
|---|---|---|
| committer | Martin Ritchie <ritchiem@apache.org> | 2009-09-28 15:33:18 +0000 |
| commit | fc1ba27a41f5a6223c7331004d1ab3a44ec07cf6 (patch) | |
| tree | 591c591d862cfa87d34284dba8424bb4e2a71b7f /cpp/src/tests/QueueTest.cpp | |
| parent | 490d068a2aeed02fb7521caccee7bff6c1b77bb7 (diff) | |
| download | qpid-python-fc1ba27a41f5a6223c7331004d1ab3a44ec07cf6.tar.gz | |
QPID-1871 : Updated RollbackOrderTest to include an onMessage test. Fixed deadlock issue with 0-10 rollback method and onMessage usage. Moved 0-10 rollback strategy to the abstract AMQSession and updated 0-8 to use that approach.
0-8 Still excluded from test runs as the race condition is not that the dispatcher would hold a message and reject after the TxRollback. The problem is the Java Broker sends a message out after the FlowOk message so the Dispatcher then sits on it, see QPID-2116. This exact problem was hidden due to the way the Dispatcher is stopped. This problem has not been addressed. The request to stop the dispatcher can only actually stop the dispatcher whilst it is in the middle of processing a message. The stop needs to interrupt the _queue.take() and hold the dispatcher AFTER the processing of any message that it needs to do: see QPID-2117.
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@819590 13f79535-47bb-0310-9956-ffa450edef68
Diffstat (limited to 'cpp/src/tests/QueueTest.cpp')
0 files changed, 0 insertions, 0 deletions