Well will add then a bit of delay and see. What u say is classical race where
at first invocation it takes a bit longer to get things up and running whereas
in second invocation case it does not and as such I miss the message right?
I would assume it would not make sense to have kind of channel "rdy" signalling
to be able to wait for it, as only sensible use for it will be in tests right?
P.
On May 19, 2011, at 4:16 PM, Martin Sustrik wrote:
> On 05/19/2011 04:11 PM, Pavel Chlupáček wrote:
>
>> socketSub.connect("ipc://foo-S")
>
> // Connecting starts at this point. We should expect it to take
> // some microseconds.
>
>> val zm = ZMQMessage("Test", serializer).sendTo(socketPub)
>
> // Message is sent immediately. Sub socket is still not connected,
> // thus the message is dropped.
>
>> assert(poller.poll(1000)> 0&& poller.pollin(idxSub), "NO message
>> in queue? ")
>
> // Poll timeouts.
>
> Martin
_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev
