[zeromq-dev] Questions about socket errors

[Brian Granger]

Hi,

I am starting to look at using the Python interface to zeromq and have
some questions about error handling.
In particular, I am wondering how socket related errors are handled in
Python.  Some examples of what I saw:

* If I do s.connect to an endpoint that doesn't exist, it returns
happily as though it were fine.
* If I do a recv/send to an endpoint that lost its connection, it
hangs or returns happliy (NOBLOCK).

I am guessing that this behavior is unique to the Python bindings.  So
first a bigger question:  what is the model
in the C API for handling these types of socket related errors?  In
particular I am interested in keeping track of what
clients are connected to an application that is listening on a set of
sockets (it is like a message hub) and knowing
when those connections are lost or error in any way.  I need to be
able to perform message routing in this application
that *depends* on what clients are connected.  At the C-level, how can
I detect and react to these errors?

The next question is how all of this is exposed in the Python bindings.

Cheers,

Brian

-- 
Brian E. Granger, Ph.D.
Assistant Professor of Physics
Cal Poly State University, San Luis Obispo
bgranger at calpoly.edu
ellisonbg at gmail.com