It's a good point. I can't think of an application that would do it on purpose. A bug in the producer might be the only scenario where it could cause problems.
I was just running some tests while preparing to upgrade to AMQ5 and wanted to be clear on this. On Tue, Feb 26, 2008 at 9:44 AM, Ben Chobot <[EMAIL PROTECTED]> wrote: > Certainly, but if order matters, wouldn't you expect to get all or none > of your messages that depend upon each other? Why would you send x as > non-persistent and x+1 as persistent if x+1 requires x to be processed? > > Roger Hoover wrote: > > There are applications in which message order matters and you generally > > wouldn't expect message properties to affect the order of delivery > unless > > you're using a selector. > > > > On Tue, Feb 26, 2008 at 8:59 AM, Ben Chobot <[EMAIL PROTECTED]> wrote: > > > > > >> Why wouldn't you want it to be this way? > >> > >> Roger Hoover wrote: > >> > >>> Using STOMP on AMQ 5, if I enqueue some persistent and non-persistent > >>> messages and then consume them, they don't get consumed in the order > in > >>> which they were produced. The non-persistent messages are delivered > >>> > >> first > >> > >>> (with their relative order preserved) followed by the persistent > >>> > >> messages > >> > >>> (also with their relative order preserved). > >>> > >>> AMQ 4 preserved message order regardless of persistence settings. > >>> > >>> Is this expected behavior? Is it controlled by any configuration? > >>> > >>> Thanks, > >>> > >>> Roger > >>> > >>> > >>> > >> > > > > > >
