to root pom.xml.
…-680) We now bundle the needed jars with the torquebox-messaging gem. The syntax for TorqueBox::Messaging::Queue/Topic.new was updated to be backwards compatible with previous versions and to match our documentation. Right now we've somewhat overloaded :naming_host and :naming_port since these should actually be the host and port of a HornetQ Netty connector (defaults to localhost:5445). Once AS7 gets remote JNDI implemented, these will once again point to a naming host/port.
This silences the huge stacktrace thrown by IronJacamar about not being able to find my mysql.jdbc.Connection.ping method which has confused several new TorqueBox users into thinking they have an error.
The queue.connect_options accessor takes a :naming_host and :naming_port options similar to what you could pass in TB 1.1 and we'll use those when grabbing the connection factory only falling back to localhost:5445 if nothing else is passed.
…ging in Rails 3.2.0
Revert "Another shot in the dark to see if this somehow broke CI." This reverts commit 8f2aa1c.
Revert "Get rid of deprecation warnings." This reverts commit 5c97255.
…e, infinispan and lucene
This version of AS7 brings in HornetQ 2.2.10 which will hopefully fix TORQUE-474 and TORQUE-592. Several deployments in Phase.STRUCTURE had their priority increased so some of ours had to be adjusted accordingly. We also can no longer set a null client_id on HornetQ Connections.
….x and padrino.
This fix does solve the connection reset issue but caused a large number of socket accept errors and the benchmark to run out of file handles under load. Likely something with the backlog or acceptor behavior needs tweaking if we queue connections.