On 2/10/07, Marshall Schor <[EMAIL PROTECTED]> wrote:
The caller checks for null, and if null, just retries. The timeout used
is the output queue "dequeue timeout" which is only used when you
are using the "chunker" queue manager which tries to reassemble
chunks of CASes in order. Otherwise it just waits forever, and if
the timer expires earlier due to notifyAll, because this is in a retry loop,
it just goes around the loop.
So the fault at this point is that setting the dequeue timeout won't have
the desired effect of timing out things in this case. (Note: the dequeue
timeout will have the effect of causing a "FINEST" log message, but
that's it, as far as I can see).
I'm not sure what's reasonable to do at this moment re: all of this...
I don't think we should o anything for this release. It seems like,
as a whole, nothing is really broken. It's just some code that could
use some cleaning up, when someone gets around to it.
-Adam
