On Tue, Aug 17, 2010 at 4:48 PM, Matt Weinstein <[email protected]> wrote:
> IMO you might consider an error return, but asserting here is fragile. > IMO waiting forever is consistent (no other recv fails because it doesn't > understand why you're doing something). The failure is easily detected and > remedied. > However, I'll make you a deal: add a setsockopt like ZMQ_SUBSCRIBE_EMPTY_OK > to remove the error behavior. Then I can happily use it if I know what I'm > doing. Here's a better deal: provide me just one real use case for blocking on a zombie SUB. "Waiting for termination" is a possible use case but using a zombie SUB for this does not make sense, it's abuse of semantics. Polling on a set of sockets that includes zombie SUBs is valid because that is precisely how you bring a zombie SUB back to life. Similarly, doing a NOBLOCK recv() on a zombie SUB is valid for the same reason. There is no defacto reason to make poll() to work like a blocking recv(). They do different things. Sorry to be difficult here but it seems a no-brainer. Zombie joke, hehe. -Pieter _______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
