Thursday September 23, 2010

[Time] NameMessage
[04:28] Pe_Ell anyone know if there is a roadmap for release dates? Just wondering since I didn't notice one on github or the site off hand...
[05:39] guido_g nope, there is no such thing
[05:40] guido_g standard quote is "it'll be released when its ready" :)
[05:42] Pe_Ell heh, k
[05:43] Pe_Ell there's a feature in 2.1 that I'm waiting for... wasn't sure how stable the git branch was... guess I'll just keep on waiting... :)
[05:43] guido_g if you're around longer, ask pieterh, mato or sustrik, they do know the details
[05:44] Pe_Ell k. I'll just leave this open for now while I read more code/docs...
[12:53] pieter_hintjens hi folk, it's quiet today
[12:54] pieter_hintjens *folks
[12:54] pieter_hintjens I've pushed out a new version of the user guide
[15:19] friendshi hi, I'm having an issue when trying to do non-block recv... I always immediately receive an exception 'Resource temporarily unavailable', anyone know where i might be going wrong?
[15:20] guido_g maybe the connect wasn't done at this point in time
[15:20] friendshi hmm it works fine with i do blocking, and i have tried sleeping for 5+ seconds between connect and recv
[15:21] guido_g because the real connection handling is in a seperate thread, you might do the receive while the underlying connection is not yet established
[15:23] friendshi i dont think that can be it... when i do it without block it begins reading my messages immediately... unless i am misunderstanding what you are suggesting
[15:23] friendshi *without nonblock
[15:24] guido_g i didn't suggest something
[15:24] friendshi sorry i just mean, what you are saying the issue could be
[15:24] guido_g one idea is to make the first receive blocking and all others not
[15:25] guido_g if this works, it's time for an issue i'd say
[15:26] friendshi same issue :(
[15:26] friendshi it does receive first message fine, then immediate fail on nonblock recv
[15:26] guido_g then try to make a small program that shows the problem and show the code
[15:27] friendshi sure
[15:45] vagn greetings. Is there an example somewhere on how to shutdown a ZMQ_REP socket so the otherside doesn't hang when the service returns?
[15:46] vagn I want to stop getting requests, but finish all the replies, then exit the program.
[15:47] vagn but the requests are queued, right? so the other side things it should get a reply. so basically I have to shutdown the incoming queue, too.
[15:47] vagn *thinks
[15:53] guido_g i'd do this kind of synchronization with explicit messages, much easier imho
[15:54] vagn thinking some more... that still doesn't cure tha case where the server crashes.
[15:56] vagn maybe I should start over and do things with pub/sub or multicast. I don't care if I lose a message. the next one that comes along is good enough.
[16:16] friendshi i find that my problem was that in pyzmq document/source it states 'If NOBLOCK is set, this method will return None if a message is not ready' but that is untrue, I believe it is throwing exception is no message exist
[16:16] friendshi maybe i am misreading or misunderstanding though
[16:22] keffo oh by my might beard! I hate 3g
[16:58] bgranger friendshi: hi