[zeromq-dev] Closed dealer socket may cause context termination to hang
Pieter Hintjens
ph at imatix.com
Mon Nov 10 22:37:58 CET 2014
That bug hasn't been fixed IMO.
So what you're saying is that simply doing a "sleep (n)" inside
zmq_ctx_term would actually work better than LINGER n seconds (where n
is less than infinity)... LOL.
Do we have a volunteer for trying this?
On Mon, Nov 10, 2014 at 5:54 PM, Frank Hartmann <soundart at gmx.net> wrote:
> Hi,
>
> Sorry to jump in into the middle. Has the LINGER option been fixed in
> the meantime (==in the last 12 month)?
>
> The only values which were working @4.0.0 as documented were 0 and
> inf. I think for all other values the system would block for the LINGER
> amount, but not send messages during that period.
>
> https://zeromq.jira.com/browse/LIBZMQ-554
>
> So basically the behaviour of a LINGER set to two seconds was at the
> time I wrote the bug:
>
> - do nothing for 2 seconds
> - stop
>
> I would have expected:
>
> - send out queued messages for 2 seconds
> - stop
>
> So clearly 0 seconds is much better if the behaviour has not been
> changed...
>
> kind regards
> Frank
>
>
>
>
>
>
> Pieter Hintjens <ph at imatix.com> writes:
>
>> Note this thread: https://github.com/zeromq/libzmq/pull/1253
>>
>> On Sat, Nov 8, 2014 at 6:38 AM, André Caron <andre.l.caron at gmail.com> wrote:
>>> Thanks for the quick replies and getting this included in the docs!
>>>
>>> Cheers,
>>>
>>> André
>>>
>>> On Thu, Nov 6, 2014 at 9:38 AM, Michi Henning <michi at triodia.com> wrote:
>>>>
>>>>
>>>> On 7 Nov 2014, at 0:33 , Pieter Hintjens <ph at imatix.com> wrote:
>>>>
>>>> > Well, I used the term "insane" in the man page.
>>>>
>>>> I was thinking of "wrong", but "insane" will do :-)
>>>>
>>>> > OK, I've sent a pull request to set LINGER to 2 seconds by default.
>>>> > If there are no complaints we can backport this to 4.1. I think the
>>>> > old behavior can be considered a bug worth fixing.
>>>>
>>>> I can't think of any realistic scenario where the old behavior would be
>>>> helpful. And, even if there are one or two applications out there that
>>>> depend on it, I think the principle of least surprise suggests that this is
>>>> a bug, not a feature.
>>>>
>>>> Michi.
>>>>
>>>> _______________________________________________
>>>> 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
>>>
>> _______________________________________________
>> 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
More information about the zeromq-dev
mailing list