[rabbitmq-discuss] Pika and Channel.Flow

[Alexis Richardson]

Allan

I won't speculate as to what Matthias is suggesting; but interestingly
there *is* a use case for ack-on-publish.

alexis


On Tue, May 25, 2010 at 6:48 PM, allan bailey <zirpubolci at gmail.com> wrote:
> Matthias,
>
> On Mon, May 24, 2010 at 10:47 PM, Matthias Radestock <matthias at rabbitmq.com>
> wrote:
>>
>> Allan,
>>
>> allan bailey wrote:
>>>
>>> On Mon, May 24, 2010 at 12:54 PM, Matthias Radestock
>>> <matthias at rabbitmq.com <mailto:matthias at rabbitmq.com>> wrote:
>>>    When the memory alarm is raised, RabbitMQ sends a channel.flow on
>>>    all channels. There are no delays other than the usual processing
>>>    and networking delays.
>>>
>>> So should I just redeclare my exchange, would that be a sufficient no-op?
>>>
>>>  I tested it just now and it works to trigger the alert.
>>
>> You shouldn't have to do send any other command - if the client library is
>> implemented correctly.
>>
>>
>
> Are you suggesting that basic_publish should return something?  From my
> brief reading of the protocol it doesn't return anything.
> In my code I'm in a tight look just sending messages via basic_publish.  At
> no point does basic_publish notice the channel.flow state change.
> I'll take a look at the java and/or php clients and see how they handle
> this.
> -allan
>
>>
>> Matthias.
>
>
> _______________________________________________
> rabbitmq-discuss mailing list
> rabbitmq-discuss at lists.rabbitmq.com
> http://lists.rabbitmq.com/cgi-bin/mailman/listinfo/rabbitmq-discuss
>
>