IRC Logs for #circuits Monday, 2017-01-16

*** re4per has joined #circuits03:59
GitHub191[circuits] prologic pushed 1 new commit to master:
GitHub191circuits/master 930ca39 Florian Apolloner: Fixed #54 -- Do not leak on_done handlers. (#195)04:50
GitHub19[circuits] prologic closed pull request #195: Fixed #54 -- Do not leak on_done handlers. (master...issue54)
GitHub107[circuits] prologic force-pushed improve_test_reliability#174 from 4795828 to cba8d08:
GitHub107circuits/improve_test_reliability#174 17a8cc7 James Mills: [tests]: Fix test_tcp_lookup failure by using an invalid FQDN04:52
GitHub107circuits/improve_test_reliability#174 d86f99e James Mills: [tests]: Fixed test_auto_reconnect node test reliability04:52
GitHub107circuits/improve_test_reliability#174 0ad61d8 James Mills: [tests]: Improve bridge and node tests for OS X04:52
GitHub188[circuits] prologic force-pushed improve_test_reliability#174 from cba8d08 to 9033b6d:
GitHub188circuits/improve_test_reliability#174 9033b6d James Mills: [tests]: Fix tests/node/
GitHub127[circuits] prologic closed pull request #176: [tests]: Improve test reliability on OS X / BSD (master...improve_test_reliability#174)
GitHub87[circuits] prologic pushed 1 new commit to master:
GitHub87circuits/master a63e641 James Mills: [tests]: Improve test reliability on OS X / BSD (#176)...04:55
*** travis-ci has joined #circuits05:00
travis-cicircuits/circuits#539 (improve_test_reliability#174 - cba8d08 : James Mills): The build has errored.05:00
travis-ciChange view :
travis-ciBuild details :
*** travis-ci has left #circuits ()05:00
*** travis-ci has joined #circuits05:04
travis-cicircuits/circuits#538 (master - 930ca39 : Florian Apolloner): The build has errored.05:04
travis-ciChange view :
travis-ciBuild details :
*** travis-ci has left #circuits ()05:04
*** travis-ci has joined #circuits05:23
travis-cicircuits/circuits#543 (master - a63e641 : James Mills): The build is still failing.05:23
travis-ciChange view :
travis-ciBuild details :
*** travis-ci has left #circuits ()05:23
*** romster_ has joined #circuits08:39
*** re4per has quit IRC10:07
*** ke4roh__ has quit IRC11:43
GitHub57[circuits] apollo13 pushed 1 new commit to master:
GitHub57circuits/master cef232f Florian Apolloner: Fixed previous commit.12:08
apollo13spaceone, prologic: any ideas on those failures: ?12:13
apollo13mhm, seems to be the result of a recent commit12:21
apollo13hihi prologic broke it with his mac fixes :รพ12:25
*** travis-ci has joined #circuits12:28
travis-cicircuits/circuits#544 (master - cef232f : Florian Apolloner): The build is still failing.12:28
travis-ciChange view :
travis-ciBuild details :
*** travis-ci has left #circuits ()12:28
*** romster_ has quit IRC12:43
apollo13pushed a fix, if someone is around I'd like to talk a bit about nodes, cause I do not really understand them yet13:08
apollo13from the test it looks as if the node and the client are in one process, which does not seem that useful?!13:09
GitHub60[circuits] apollo13 closed pull request #194: Removed python 2.6 and 3.x < 3.4 (master...python_version)
GitHub107[circuits] apollo13 pushed 2 new commits to master:
GitHub107circuits/master 9cf7f8a Florian Apolloner: Removed python 2.6 and 3.x < 3.413:09
GitHub107circuits/master 780d357 Florian Apolloner: Small fixes.13:09
*** ke4roh__ has joined #circuits13:12
GitHub148[circuits] apollo13 pushed 1 new commit to master:
GitHub148circuits/master b4f16db Florian Apolloner: Fixed auto_reconnect test.13:15
*** travis-ci has joined #circuits13:39
travis-cicircuits/circuits#546 (master - 780d357 : Florian Apolloner): The build has errored.13:39
travis-ciChange view :
travis-ciBuild details :
*** travis-ci has left #circuits ()13:39
*** travis-ci has joined #circuits13:51
travis-cicircuits/circuits#547 (master - b4f16db : Florian Apolloner): The build has errored.13:51
travis-ciChange view :
travis-ciBuild details :
*** travis-ci has left #circuits ()13:51
apollo13okay, someone has to explain to me how the watcher in the tests works13:52
apollo13to the best I can tell all this is single threaded13:52
rioti may have asked in the past, but has anyone integrated zmq into circuits?14:58
apollo13riot: should be a simple component, no?14:58
apollo13I mean zmq uses it's own threadpool anyways and shouldn't block (tm) IIRC14:58
riottried adding some sample code for a peer2peer chat. First thing it did was blocking. Tried wrapping it via worker pool and it just doesn't seem to do anything anymore :/15:00
rioti'm pretty sure, the worker pool wasn't meant for this. Will probably have to rethink the whole situation about p2p. Meh.15:01
riotthat is the one i tried converting:
riot(it is actually very near my actual usecase for zmq)15:02
rioti'd just have to recheck for new listeners every now and then, as i expect them to suddenly (dis-)appear in the mesh network.15:04
apollo13riot: ah, I was wrong:
apollo13you need a poller it seems15:04
apollo13sadly zmq does not expose posix fds, which makes integration into another eventloop somewhat annoying15:05
riotwell, that seems like one way to do it. Did you check out the example i used? Another way (more like what i need ;)15:06
apollo13by putting it into a thread?15:07
apollo13yeah that will probably work too15:07
apollo13you could also try to adapt
apollo13prologic: are you around? :D15:39
riothmm, more of this ZMQStream/Eventloop business:
*** ke4roh_ has joined #circuits15:58
riot << looks simple enough. I'll try adapting that.15:58
*** ke4roh__ has quit IRC16:00
riothm, running loop.start() blocks (obviously)16:12
riotcould this be of use: ?16:12
apollo13riot: yeah, but you'd have to see on how to replace the circuits loop I think16:31
apollo13riot: the main issue here is that zmq has no posix file descriptors, hence you cannot just register it in an existing mainloop16:40
riotso i need another way :/16:50
riotwrapping the whole ioloop bunch into a thread -> no help - it opens the socket but i see no incoming data at all17:10
rioti assume a context problem or something. Annoying.17:10
*** ke4roh__ has joined #circuits17:12
*** ke4roh_ has quit IRC17:14
*** Coldblackice has joined #circuits19:07
riotoi, i was testing wrongly, zmq + telnet = nothing19:35
rioti threaded it and it works19:35
prologicapollo13: hey20:55
prologicI'm sort of aroud but really tired / jet lagged20:55
prologicre watcher in tests; it's more of just an event capture/logger and helper to assert an event was seen20:56
prologicriot: the way I would do zmq integration is like what I did with -- if you want tight integration20:57
prologicin autodock I integrate and translate docker events into circuits equivalent so that I ca do distributed event'ing via a custom circuits.node implementation20:57
apollo13prologic: ok, will catch you once you are back fully -- not in a rush :D21:47
riotprologic: nah, i don't want too tight integration - essentially a matter of security, so i think, having one central component dealing with the insecure side makes much sense21:57
riotalso, i want to build a similar piece of infrastructure to deliver mqtt data21:58
*** ke4roh__ has quit IRC22:30

Generated by 2.14.0 by Marius Gedminas - find it at!