I noticed there has been a lot of traffic about this particular function and the potential for handle leaks, but no one has asked why this function uses a fixed port instead of a dynamic port to establish the pair. I recently encountered a problem where port 5905 was in use, and it causes an exception that kills the application (on this particular machine we increased the dynamic port range to include 5905).
Is there some reason why the port is hardcoded instead of simply using whatever port is available? It seems an oversight to me that I may have to recompile ZeroMQ to use a different port if our software is going to be installed on a device where that port is already taken. I notice in the comments of the function that it mentions using a dynamic port, but doesn't actually do so. Stew Scott
_______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
