Wednesday April 27, 2011

[Time] NameMessage
[00:09] zedas ah ok it works but you have to wait a bit for the clients to reconnect
[07:56] pieterh hi mikko
[07:57] guido_g good morning ømq
[08:00] pieterh hi guido_g!
[08:00] guido_g good morning pieterh
[08:15] mikko pieterh: hi
[08:15] pieterh hi mikko!
[08:15] pieterh you asked if people were arriving on the 9th
[08:15] pieterh some are, yes
[08:15] mikko pieterh: i'll book tickets 9th May - 10t Mayh
[08:16] pieterh cool!
[08:16] mikko coming back from holidays today
[08:16] pieterh I think it's sane to hold a meetup on the evening of the 9th
[08:16] mikko can i stay at yours?
[08:16] mikko or should i book a hotel?
[08:16] pieterh for sure
[08:16] pieterh stay with us, mato will also be here
[08:16] mikko thanks
[08:17] pieterh lots of space, maybe bring a sleeping bag if you have one
[08:17] mikko i can buy one, it probably has use later
[08:17] mikko i gotta run, check out soon
[08:17] pieterh nah, don't waste money for that
[08:17] pieterh cyl
[08:17] mikko ill be back around GMT 22:00
[08:17] pieterh :)
[08:17] pieterh I'll be drumming
[08:20] guido_g pieterh: you're writing an app? like app in "app for a phone"?
[08:20] pieterh hmm, yes
[08:20] guido_g just saw that tweet of yours
[08:20] guido_g ahh i see
[08:20] pieterh its a prototype for a new mobile concept
[08:20] pieterh has nothing to do with zeromq sadly
[08:21] guido_g hehe
[08:21] pieterh except perhaps to convince me we need an HTTP bridge sooner rather than later
[08:21] guido_g or real operating systems for mobile devices
[08:22] pieterh as compared to Linux? :)
[08:22] guido_g no, of course not
[08:23] pieterh The problem isn't really running the 0MQ stack on Android or iOS
[08:23] guido_g but this tiny things like s60 or ios will not last any longer
[08:23] pieterh for sure
[08:23] guido_g as ia said, real oses
[08:23] guido_g omg
[08:23] pieterh there is only one real os, eventually
[08:23] guido_g let me type this again: as i said
[08:24] guido_g at only one open source
[08:24] pieterh the obvious sometimes takes a long time to happen
[08:24] guido_g i don't know how close the windows mobile os is to the real windows
[08:25] pieterh real windows is barely a real OS by modern standards
[08:25] guido_g *sigh*
[08:25] guido_g but we do have a build process for it :)
[08:25] guido_g we := ømq
[08:25] pieterh indeed
[08:25] pieterh problem isn't building 0MQ on random mobile device
[08:26] pieterh they're all real computers with what approaches an OS
[08:26] Guthur windows 8 will probably be that 'real OS' on a mobile
[08:26] pieterh problem is running unsecured protocols over the Internet
[08:26] Guthur it will have support for the ARM arch
[08:26] guido_g Guthur: so there two then, and one if them is free and (nearly) open source :)
[08:27] Guthur guido_g: Android?
[08:27] guido_g Guthur: jepp
[08:27] guido_g i consider my beloved maemo dead
[08:27] Guthur Google got slapped with a fine recently in Texas, for patent infringement related to the linux kernel
[08:28] guido_g and meego is a still born
[08:28] Guthur $5m
[08:28] guido_g yeah, read about it
[08:29] pieterh Google are big enough to actually take on the patent system, when they get annoyed enough
[08:29] Guthur this patent is so bogus, it might actually wake people up to the fact it is very broken
[08:30] pieterh Guthur: you think if it wasn't a bogus patent, that'd be better somehow? :)
[08:30] Guthur well...
[08:31] pieterh Imagine some firm comes up with a really valid patent for 'sockets that automatically fan in and fan out'
[08:31] pieterh filed in 1999 or somesuch
[08:31] Guthur did you file it, hehe
[08:31] pieterh sold to some random holding firm... never turned into software...
[08:31] pieterh no, I'm pointing out that it's not about good vs. bogus patents
[08:32] guido_g like software and idea patents
[08:32] Guthur hehe yeah I've read on the net that your have been quite vocal in the past against the current system
[08:33] pieterh if you want a history lesson in all that is bad about patents:
[08:33] pieterh this is one reason 0MQ is not a conventional commercial closed-source product
[09:09] djc pieterh: as for http bridge, I have some code that does 0MQ-to-WebSockets
[09:09] pieterh djc: that's interesting, have you published it?
[09:09] djc not yet
[09:10] djc it's written in Python and it contains some idiosyncrasies specific to the content we're sending over 0MQ
[09:10] djc but I could probably publish it somewhere
[09:10] pieterh so its not a generic bridge
[09:10] djc most of it is generic
[09:10] djc but we're sending binary content over zeromq, so you need something to translate to JSON (at least, that seemed easier than parsing the binary stuff in JS)
[09:11] pieterh for sure
[09:11] pieterh what version of websockets?
[09:11] djc whatever Fx4 and currently stable Chrome support
[09:11] pieterh ok, the old hixie spec
[09:11] djc (you need to tweak a flag in Fx4 because this is the not-so-secure version)
[09:12] djc we're actually using this in production, but of course I'll update it as soon as the browsers implement the new stuff
[09:12] pieterh there's a topic page on the wiki, it'd be useful to post something about your websockets bridge IMO
[09:12] guido_g q: isn't websockets just plain tcp after the http upgrade?
[09:12] pieterh guido_g: hah
[09:12] djc guido_g: it's not
[09:12] djc pieterh: where's that at?
[09:13] guido_g i'm puzzled
[09:13] pieterh yeah
[09:13] pieterh guido_g: it does, at the least, framing
[09:13] pieterh when it went through the ietf mangler it got a lot more complex
[09:14] guido_g ok
[09:14] guido_g will look it up later then
[09:14] guido_g really thought that after the upgrade the os socket be used as a plain tcp socket
[09:15] djc no
[09:15] djc messages are delineated by one-byte header and footer
[09:15] djc so it's a stream of messages, not a stream of bytes
[09:15] djc which is probably easier to handle in JS and also maps quite well to 0MQ
[09:15] guido_g i do not dare to ask why :)
[09:19] djc ugh, the wikidot thing confuses me
[09:20] djc I made an account and am trying to comment on the 0mq-over-http page, but it keeps saying "or Sign in as Wikidot user", even though I activated my account and signed in
[09:21] guido_g cookies
[09:21] guido_g incl. third party ones :(
[09:22] djc I deleted all cookies on zeromq-containing domains and the same thing still happens
[09:23] guido_g you've to allow cookies to be set
[09:29] pieterh djc: ... are you behind some firewall or proxy?
[09:31] djc nope
[09:31] djc there, wrote an account-less comment
[09:37] pieterh djc: what browser are you using?
[09:39] djc Fx4
[09:42] pieterh Fx? Firefox?
[09:42] pieterh so it's doing something (paranoid/silly/too clever) with cookies
[09:53] guido_g firefox 4 works, just logged in
[11:13] pieterh sustrik: ping?
[18:53] Dantman Are there any api docs for jzmq?
[19:34] staylor is there a way to get a source ip address of a socket in zeromq given it's id? DEALER->ROUTER where router side requests the id of DEALER
[21:10] Guthur Dantman, there is some example code in the guide
[21:10] Guthur staylor, not that I am aware of
[21:11] Dantman I know there is the guide, some example code, and zmq docs... but it's a little awkward to read zmq docs to look at the api and try to translate that to Java
[21:11] Dantman ...especially since I don't do C
[21:54] mikko good evening
[21:59] mikko pieterh: eurostar booked
[22:20] staylor I asked earlier, wondering if anyone knows if it's possible to get the remote ip address of a socket? ie: a router socket getting the ip address of the dealer socket connected to it?
[22:26] mikko staylor: nope, you have to embed this information into messages if you need it
[22:44] staylor mikko: yeah encoding it into the message isn't really possible since the dealer could be behind a nat