[rabbitmq-discuss] Socket leak in java client library

[Steve Powell]

Dear Iain,

Looking at the newConnection() code in detail, I have chosen not to do connection cleanup in newConnection() (as in your patch) but in AMQConnection.start() where (most of) it belongs.   In other words, when AMQConnection.start() detects a problem, it cleans up any debris (threads, heartbeat executors and/or sockets) it may have created during socket setup and connection negotiations before throwing an exception.  This allows us to ignore errors (except for reporting) in the newConnection loop without leaking resources.

If this passes QA it should appear in the next release, and be documented in the JavaDoc. These fixes are presently under the bug24597 track.

Steve Powell  (a tired bunny)
----------some more definitions from the SPD----------
avoirdupois (phr.) 'Would you like peas with that?'
distribute (v.) To denigrate an award ceremony.
definite (phr.) 'It's hard of hearing, I think.'
modest (n.) The most mod.

On 28 Nov 2011, at 20:39, Iain Hull wrote:
> http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/2011-November/016612.html