On Wed, May 18, 2011 at 9:44 AM, John D. Mitchell <[email protected]> wrote:

> The observed behavior seems to be that the new connections are *not* 
> completed while the  publisher is busy spewing as fast as it's able to and so 
> the latecomers are basically ignored until/unless there's a gap for them to 
> sneak through.

This is possible.

There is nothing blatantly obvious in your code, it's a bit more
complex than I'd make (e.g. you don't need multipart handling for the
synchronization dialog) but it looks accurate.

If you can reliably reproduce unfairness in the scheduling, log an
issue with a nice minimal test case, and we'll handle it. I'd suggest
trying to make a single-process test case that starts two threads, and
prints a good "FAILED" message when the test fails, that makes life
easier for everyone (except you, such elegant test cases are quite
hard work. :)

-Pieter
_______________________________________________
zeromq-dev mailing list
[email protected]
http://lists.zeromq.org/mailman/listinfo/zeromq-dev

Reply via email to