[zeromq-dev] HWM behaviour & blocking

[Steffen Mueller]

Hi Pieter,

On 05/10/2012 02:36 PM, Pieter Hintjens wrote:
> On Thu, May 10, 2012 at 7:42 AM, Steffen Mueller
> <zeromq at steffen-mueller.net>  wrote:
>
>> Frankly, I fail to see what's so funky about not wanting to block the sender
>> if the receiver is restarted, using the queuing of the library[1].

[snip, explanation of why I was looking for the functionality]

>> Either way, I understand the message (and Chuck's, too). I'll roll my own
>> solution.
>
> You seem annoyed that ZeroMQ somehow does not live up to your
> expectations. Yet the patterns which ZeroMQ enables are well
> documented, and the material that explains how to build on top is
> vast, and translated into dozens of programming languages.

Oh, I wasn't annoyed at the library. If my tone was sour, I apologize!

I think at the same time, I wasn't clear enough on what I meant by "the 
message" and "roll my own". I meant that the message had been "0MQ is 
not meant to do that and won't" and the conclusion was that I "need to 
roll my own on top of 0MQ", not that I intended to "ditch 0MQ and start 
from TCP".

> I've no strong opinion on this, but you might reflect on what it looks
> like to others. You find a free library, made by others over years at
> their expense. You do not read the available material (or you skim
> it). You ask for, and get expert advice, for free. Then you complain
> that the tool doesn't fit your personal use case as though the
> universe was designed for you?
>
> Steffen, seriously? Learn it, use it, and if you can improve it, send
> us a patch.

Don't be condescending. I've written and supported a lot of free 
software myself. That's not to say anything would entitle me to act like 
a dick. Again, if that was the impression I created, I'm really sorry. 
This logic goes both ways.

--Steffen