Re: [Mod-pubsub-developer] Non-blocking publishing in Python

[<mod...@li...>]

A:

> xmlBlaster and Activ MQ's HTTP for asynchronous notification either
> require polling or a proxy that can accept connections across the
> firewall. The beauty (and scariness) of mod-pubsub's approach is the
> use of streaming HTTP to keep connections persistent so no polling and
> no proxy needed.


At what level of granularity is polling not required? The reason I ask is that
the Python pubsublib client very much uses polling:

    while 1:
        asyncore.poll(scheduler.timeout())
        # print "\n\n asyncore.poll \n"
        scheduler.run()

Going back to the original question I had, this is the core of the problem since
the while loop is very much blocking and is processor intensive. From the
original Medusa documentation on asyncore, I offer the following quote:

"Now that you know what select() does, you're ready for the final piece of the
puzzle: the main polling loop. This is nothing more than a simple while loop
that continually calls select() with a timeout (I usually use a 30-second
timeout). Such a program will use virtually no CPU if your server is idle; it
spends most of its time letting the operating system do the waiting for it.
This is much more efficient than a busy-wait loop."

Hence the solution was to do just this: replace the while loop with the loop()
command which is much more CPU-friendly.

  asyncore.loop()

Anyone interested in embedding the client in a larger python project will likely
need to do this. Ultimately, this is a more efficient implementation and should
probably be merged into the code base. Do you want a patch?

Brian

-- 
  ====
brian lee dae yung
biomimetic art and research
  mux space . com

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.