[zeromq-dev] Apparent queuing on PUB/SUB fan-in despite HWM=1

Brian Rossa rossabri at hotmail.com
Fri May 13 20:02:17 CEST 2011

> >> Can you post a minimal Python test case, please?
> > I've eschewed minimalism for realism here. I don't think it will be hard for
> > you guys to reproduce my results, though; I've been tight with the code.
> Brian, there are good reasons why we ask for a minimal test case. If
> there is indeed an issue with pub/sub and hwm=1, then you can
> demonstrate that in 10 lines of Python. Often, trying to reduce a
> "realistic" example to a minimal one will show you where the problem
> is.
> Please do create a minimal test case that proves the issue, and which
> we can then translate into C as we need to. Otherwise we'll conclude
> it's something in your code.

I understand that you guys needs something small and tractable that you can put in your pocket, take back to the shop, and roll into the libraries. I'll do my best!
I really feel that the BIG PICTURE question I want to answer can't be boiled down to a unit test so easily, though. The Question: "Is ZeroMQ suitable for use in real-time and just-in-time systems?" I think my approach, up to and including the test scripts I've sent, is pretty reasonable and I hope you'll have a look at it when you can. It has identified at least one use-case -- PUB/SUB fan-in -- in which the answer seems to be "no." I am open to suggestions about how to turn those BIG PICTURE results into a constructive contribution to the project.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.zeromq.org/pipermail/zeromq-dev/attachments/20110513/1c849178/attachment.htm>

More information about the zeromq-dev mailing list