On Wed, 2014-08-13 at 18:41 +0100, CLIVE wrote: > The patch I applied at work to solve this issue has been working for a > while, I'm just struggling with time at the moment to get it > incorporated in to the QPID code base. > > All I did was add a 'reset()' method to the session and then used the > functionality already present in the qpid::messaging::Client::reconnect > method to create a new session and loop round recreating any associated > senders and receivers. > > If someone else wants to pick up the JIRA and apply the change then > please feel free, otherwise I will try and get it applied on my next day > off.
Is this the right JIRA? https://issues.apache.org/jira/browse/QPID-5824 It is labeled "patch" but there's no attachment. If you attach your patch I can have a look. > > Clive > > > On 12/08/2014 14:38, Gordon Sim wrote: > > On 08/12/2014 02:28 PM, Jakub Scholz wrote: > >> In the qpid.messaging C++ API, many exceptions don't close the > >> connection > >> but only the session - for example when one tries to create a > >> receiver on a > >> queue which doesn't exist. > >> > >> As far as I understood it, in such situation I have to always > >> recreate the > >> session and all its senders / receivers manually one by one in my > >> application, is that correct? > > > > Yes, at present that is correct. > > > > (With AMQP 1.0, errors tend to be signalled at the link level, whereas > > with 0-10 it was at the session level). > > > >> Or is there some way how to let the library > >> recreate all of them for me - e.g. something like the reconnect() > >> method on > >> the connection level? > > > > There was a JIRA created[1] by Clive Lilley who I believe had a patch > > for this in the works. Clive, if you are reading this, any update? Is > > it still something you are working on? > > > > [1] https://issues.apache.org/jira/browse/QPID-5824 > > > > > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [email protected] > > For additional commands, e-mail: [email protected] > > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
