[zeromq-dev] EAGAIN that I can't explain on zmq_msg_send to TCP DEALER

[Benson Margulies]

On Thu, Apr 2, 2015 at 5:12 AM, Pieter Hintjens <ph at imatix.com> wrote:
> In general if you have such questions, providing a minimal example is
> the best way to get an answer.

Fair enough. Before undertaking the development of that, I wanted to
check whether I had missed a documented condition in which this would
be an expected result. I'm going to read you as confirming my
understanding of the doc and start marshalling an example and see if I
can set up the failure.


>
> On Tue, Mar 24, 2015 at 2:03 PM, Benson Margulies <bimargulies at gmail.com> wrote:
>> Linux. ZMQ 4.0.5.
>>
>> I have a thread; it reads from a PULL (inproc) and writes to a DEALER
>> (tcp). It is driven in a loop with zmq_poll, in fact, it multiplexes
>> several PULL sockets to the one DEALER.
>>
>> It reads with zmq_msg_recv and sends with zmq_msg_send, recycling the
>> same msg over and over.
>>
>> It never passes the nonblocking flag. I set no explicit HWM or other
>> socket options.
>>
>> Sooner or later, zmq_msg_send fails with EAGAIN on the DEALER.
>>
>> As I read the doc, this should not be possible.
>>
>> (Note: i thought I sent a version of this question before, but I do
>> not see it in the archive and I never got a reply. apologies if it
>> came through and is somehow beneath notice.)
>> _______________________________________________
>> zeromq-dev mailing list
>> zeromq-dev at lists.zeromq.org
>> http://lists.zeromq.org/mailman/listinfo/zeromq-dev
> _______________________________________________
> zeromq-dev mailing list
> zeromq-dev at lists.zeromq.org
> http://lists.zeromq.org/mailman/listinfo/zeromq-dev