Skip navigation

I2P Development Meeting 211

I2P Development Meeting 211

I2P dev meeting, December 4, 2012 @ 20:00 UTC

Quick recap

  • Present:

    dg, hottuna, KillYourTV, lillith, Meeh, psi, str4d, weltende, zzz

  • Next Meeting

    The next meeting is scheduled for Tuesday, December 11 @ 20:00 UTC (8:00PM)

Full IRC Log

20:18:53  * KillYourTV has noticed that we're 17 minutes into the meeting...and we're off to a quiet start...
20:19:31  <lillith> i was wondering that, did i also get the wrong time or something?
20:20:23  * dg is waiting for self to be free
20:20:30  <dg> some stuff needs wrapping up first
20:20:33  <dg> sorry
20:20:39  <dg> you guys can start without me if you need to
20:23:07  * KillYourTV 's mostly going to be an observer due to his rather insignificant & unimportant roles...
20:23:15  <KillYourTV> ...so start times matter not.
20:23:39  <Meeh> I will be here, but I can wait until dg is ready
20:23:42  <str4d> I wonder if it would be possible to have two IRC leaf connections to the same leaf to reduce netsplits...
20:23:48  <dg> As long as nobody's becoming annoyed, I'll wait a little longer
20:23:59  <dg> Hopefully $task won't be too much longer
20:24:02  <str4d> (without doubling up on messages)
20:24:05  <dg> str4d: not without srs mods to IRCd
20:24:16  <dg> (or i2p hax?)
20:24:22  * KillYourTV nices his tasks to give more CPU time to dg
20:25:07  <str4d> There are already I2P mods for the IRCd so why not redundancy?
20:25:50  <dg> actually
20:25:50  <str4d> I guess it depends on the IRCd
20:26:04  <dg> I kind of see the amount of IRC splits as a way of measuring network health
20:26:19  <dg> For me, it says something about tunnel success :-P
20:27:07  <str4d> Speaking of which.
20:27:30  <KillYourTV> I don't know what mods were applied and why they were needed. (Back in the day ngircd needed a brief mod for b32 displaying...but with webirc it's not needed (and they're not displayed here anyway))
20:28:00  <str4d> -_-
20:49:54  <psi> orion: re: c++ i2p you mind if i add a build system to your code? probably scons
20:54:16  <dg> I'm ready
20:54:27  <dg> Sorry for the delay, folks
20:55:01  * dg pings #i2p-dev
21:03:16  <str4d> o/ dg
21:04:09  <hottuna> 'lo dg
21:05:07  <iRelay> <weltende@freenode> *waves*
21:05:53  <psi> yay
21:05:56  * psi timed out
21:11:17  <hottuna> ready dg?
21:13:23  <dg> sorry, I d/c'd
21:13:26  <dg> What did I miss?
21:13:26  <dg> <+iRelay> <weltende@freenode> *waves*
21:13:26  <dg> »» chanserv gives voice to psi
21:13:26  <dg> <+psi> yay
21:13:26  <dg> »» +psi timed out
21:13:26  <dg> »» s-771 is now known as s-77
21:13:26  <dg> <+dg> yay!
21:13:26  <dg> <+dg> Everyone woke up for me<3
21:13:27  <dg> <+dg> weltende: Any news regarding mailing list? plz have some for me
21:13:27  <Meeh> *waves*
21:13:27  <hottuna> I dont think you missed anything
21:13:27  <psi> yup
21:13:27  <dg> alrighty
21:13:27  <dg> So, weltende: ..
21:14:37  <iRelay> <weltende@freenode> not really.. no time so far
21:14:47  <iRelay> <weltende@freenode> kytv could do it.. he has root access to the box I had in mind afair ;-)
21:14:50  * dg waits a few minutes
21:15:41  * dg pokes KillYourTV
21:17:30  <iRelay> <weltende@freenode> dg: and more cowbell!
21:17:34  <iRelay> <weltende@freenode> err.. structure
21:18:02  * dg was never any good at that, but he'll try
21:18:12  * dg moves on
21:18:15  <dg> KillYourTV: ping when back
21:18:21  <dg> So, status updates.
21:18:40  * str4d has ~40 mins, so GTFG fg ;P
21:18:40  <KillYourTV> hmm? I didn't realize I had root but I can to throw some time at it (coursework & moving has taken up a considerable bit of time)
21:18:40  <dg> Are they worth bringing back since we have the meetings (and hopefully the summaries of them, but they're not working out right now)?
21:19:15  <str4d> dg, who would the status updates be for?
21:19:22  <str4d> s/for/aimed at/
21:19:25  <iRelay> str4d meant: dg, who would the status updates be aimed at?
21:19:34  <dg> Well, I assume the previous ones which jrandom maintained were for those who wanted a TL;DR of the meetings
21:20:06  <dg> Perhaps what we did with the last meeting (on the website) where the results were summarized at the top of the page
21:20:31  <hottuna> They would be nice for the sake of community-building, but also a bit of work.
21:20:49  * dg doesn't think it's a necessity as such but it'd be nice to say what the meeting accomplished/decided
21:21:32  <hottuna> Would anyone like to do it?
21:22:01  <str4d> On the site revamp I've taken the status updates as "blog entries"
21:22:05  <str4d> So it might be better to write those, or structure the status updates as such.
21:22:20  <hottuna> that sounds pretty good
21:22:57  <str4d> (And since the blog will end up with RSS or whatever, that can then be used to feed into any other distribution lines that are desired)
21:23:24  <psi> point 1: to consider, CCC
21:23:27  <lillith> i was thinking after a meeting we could have a new thread on zzz.i2p, to allow people to reply with any follow up things from what has been discussed
21:23:30  <psi> how will "this side" organize?
21:23:55  <hottuna> I support the blog idea, however someone would have to do it.
21:24:50  <psi> oh damn lag
21:25:27  <str4d> And the status updates don't need to necessarily be too "minutes-y", since the meetings section of the revamp is where actual minutes should go (and I'm thinking that the minutes could be put into a feed as well, while the full logs are displayed with the minutes on the site)
21:26:10  <zzz> dg, fyi, jr's status updates were sent out hours before each meeting, and they were not minutes of the previous meeting
21:26:55  <str4d> Ah, thanks zzz - so more of a general get-everyone-up-to-speed-before-the-meeting update.
21:27:10  <zzz> correct
21:27:39  <zzz> dg, fyi, jr's status updates were sent out hours before each meeting, and they were not minutes of the previous meeting
21:28:00  <dg> zzz: ah ok, it was a tl;dr of the $week?
21:28:03  <dg> <+dg> Moving on?
21:28:03  <dg> <+dg> psi: ccc is on the agenda :)
21:28:03  <dg> ---» aquarium (grenze@irc2p) has joined #i2p-dev
21:28:06  <dg> «--- w8rabbit (w8rabbit@irc2p) has quit (Killed (nickserv (Nick kill enforced)))
21:28:06  <zzz> <str4d> Ah, thanks zzz - so more of a general get-everyone-up-to-speed-before-the-meeting update.
21:28:06  <zzz> <zzz> correct
21:28:09  <dg> <+dg> Next topic: * PR management role (http://zzz.i2p/topics/1299)
21:28:09  <dg> <+dg> I don't know if the guy who posted that is here..
21:28:09  <str4d> dg, repost: And the status updates don't need to necessarily be too "minutes-y", since the meetings section of the revamp is where actual minutes should go (and I'm thinking that the minutes could be put into a feed as well, while the full logs are displayed with the minutes on the site)
21:28:24  <str4d> Yes he is - orion?
21:28:34  <hottuna> dg, would you be willing to write a status update before meetings in the blog?
21:29:16  <str4d> (he's in-chan at least)
21:29:23  <str4d> And FTR this is the guy working on I2PCPP
21:30:15  * psi is compiling i2pcpp
21:31:12  <psi> i am also looking at the code as well
21:32:58  <psi> i've got a SConstruct file
21:36:03  <dg> damn rats eating the cables
21:36:03  <dg> [repost]
21:36:03  <dg> <+psi> i've got a SConstruct file
21:36:05  <dg> <+dg> str4d:
21:36:05  <dg> <+dg> * Website revamp updates
21:36:05  <dg> <+dg> Anything?
21:36:08  <dg> [/repost]
21:36:12  <str4d> dg, need m0ar stables
21:36:15  <str4d> =P
21:36:21  <dg> :(
21:36:36  <str4d> dg, I've got per-net urls working
21:36:59  <dg> oh nice, progress
21:37:02  <dg> How did you do it?
21:37:06  * psi note to self don't compile with -j8 on a machine with 4 cores
21:37:13  <str4d> So in the page files, if you put <a href="http://{{ i2pconv(trac.i2p2.i2p) }}/"> it will convert to trac.i2p2.de
21:37:32  <str4d> And likewise for any other sites that have (hardcoded) known public urls.
21:37:39  <str4d> Otherwise it appends .to
21:38:16  <str4d> dg, problem was that Flask was caching filters applied to strings.
21:38:19  <str4d> So it was eval-ed on first template read and then stored.
21:38:39  <iRelay> <weltende@freenode> psi: unless it does multiple threads per core ;)
21:38:43  <str4d> Turning the filter into a context processor (so the func is eval-ed on every request) did the job.
21:38:46  <dg> Oh, ha
21:39:34  <str4d> I can try generalize the function so you pass in the entire URL and it finds and changes the domain bit, if people would prefer to use it that way.
21:39:45  <str4d> But it Works For Now (TM)
21:40:01  <dg> psi:
21:40:04  <dg> * CCC workshop/lightning talk discussion
21:40:28  <psi> yes
21:40:39  * psi reviews zzz.i2p link
21:40:58  <str4d> (aside: Once I get some free time (after getting the bugs out of the feed mechanism) I'd like to work out the download mirroring stuff with welterde.)
21:42:36  <iRelay> <weltende@freenode> str4d: well.. should be simple enough.. text file in mtn with list of all http, ftp mirrors..
21:44:32  <iRelay> <weltende@freenode> (and in the backend just an rsync master site, from which all mirrors pull)
21:44:40  <hottuna> did we choose a lightning talk topic?
21:44:40  <psi> ok regarding CCC there is 0% chance for me to get the required docs to get there in time
21:44:40  <psi> also... lots of "other stuff"
21:44:40  <psi> in general I am overloaded due to finals
21:44:40  <psi> also lag
21:45:31  <str4d> welterde, I guessed so, but I'm not familiar with the current mirror setup.
21:47:26  <str4d> The other thing, of course, is migrating the rest of the old pages across (and tidying up nav layout)
21:48:06  <dg> <+psi> ok regarding CCC there is 0% chance for me to get the required docs to get there in time
21:48:06  <dg> <+psi> also... lots of "other stuff"
21:48:06  <dg> <+psi> in general I am overloaded due to finals
21:48:06  <dg> <+psi> also lag
21:48:06  <dg> <+dg> Could probably ask the audience who has used i2p before
21:48:09  <dg> <+iRelay> <weltende@freenode> (and in the backend just an rsync master site, from which all mirrors pull)
21:48:09  <dg> <+dg> psi: yeah, ech and welt are going though afaik
21:48:11  <dg> <+str4d> welterde, I guessed so, but I'm not familiar with the current mirror setup.
21:51:57  <str4d> Okay, heading off o/
21:52:13  <dg> bye o/
21:52:28  <dg> We really should have started earlier
21:52:31  <dg> g'damnit
21:52:47  <str4d> I'll see if I can be back in time for the end, but no guarantees.
21:53:35  <dg> alright, stenography
21:53:42  * dg pokes Meeh
21:54:01  <psi> we need a generic interface for making transports
21:54:04  <psi> (imo)
21:54:23  <psi> s/need/should\ have/
21:54:26  <iRelay> psi meant: we should\ have a generic interface for making transports
21:55:08  <Meeh> I'm here, sorry just got disturbed with a phone call, back now
21:55:16  <dg> afaik there's something called "restricted routes" but I don't know how they work
21:55:19  <Meeh> *catchin up/reading log*
21:55:22  <dg> (nor have they been implemented..?)
21:55:28  <dg> Meeh: there isn't too much to read.. :(
21:55:50  <psi> dg no docs on that?
21:56:10  <dg> psi: mention on www.i2p2.i2p is all I found under roadmap or something..
21:56:47  <psi> if anyone happens to remember what "restricted routes" are/were please speak up
21:56:50  <lillith> dg, as i understand it restricted routes are like 'darknet mode' on freenet, you only connect via trusted peers
21:56:57  <psi> ah
21:57:16  <dg> ah
21:57:31  <lillith> i think :)
21:57:34  <psi> sounds like that could be it
21:58:11  <dg> matches the name
21:58:33  <lillith> it's been mentioned on zzz.i2p recently iirc
21:59:40  <psi> if someone who does know for sure from way back when a "confirmation" would be great
22:01:31  <lillith> http://zzz.i2p/topics/114
22:04:31  * dg reads
22:04:31  <lillith> it's not what i meant, but it explains pretty thoroughly
22:06:02  <Meeh> just wondering, where are we in the meeting?
22:06:13  <Meeh> what's current topic
22:06:16  <dg> We're kind of floating around, Meeh
22:06:23  <Meeh> ah ok
22:06:31  <dg> "* Hide I2P traffic. Like Tor, hide so it looks like SSL traffic, or something. (Considering countries where darknets is illegal) "
22:06:41  <dg> (We started late and thus sucking)
22:08:10  <Meeh> yea, we should think about possible alternative transport for countries blocking and making darknets illegal
22:09:47  <lillith> well to start with: how distinctive is i2p traffic now?
22:09:50  <hottuna> I think we ought to play something like that the same way tor does
22:09:50  <hottuna> and deploy it not before it is needed
22:09:50  <hottuna> as to prolong any arms-race
22:09:50  <hottuna> but we haven't been blocked anywhere yet
22:09:50  <hottuna> as far as I know
22:10:05  <dg> (yet)
22:10:25  <dg> Also, the lack of this sort of "protection" i.e system keeps some away from i2p
22:10:32  <psi> in general a generic transport api would be a developer's goldmine
22:10:32  <hottuna> we have a lot of random data, but none of the headers of ssl
22:10:57  <dg> obfsproxy is amazing but we don't need something of that level
22:11:48  <psi> some interface that you'd implement that does your version of data transport wether it's over goats or http+ssl
22:11:50  <iRelay> <weltende@freenode> psi: we already have a generic transport api ;)
22:11:54  <lillith> i2p can't really be 'too secure'
22:12:02  <hottuna> a transport api would be a good idea, and would allow for rapid development of needed transports
22:12:14  <psi> we do eh?
22:12:36  <psi> i need to look at the code closer
22:12:56  <psi> either it's not sticking out well or i over looked it or it's not there
22:13:03  <Meeh> yea, yet.. it's a matter of time
22:13:36  <iRelay> <weltende@freenode> router/java/src/net/i2p/router/transport/Transport.java is the interface you have implement
22:13:36  <dg> certainly not, lillith
22:13:46  <dg> although i2p is already a lot of crypto
22:14:17  <hottuna> i think obfsproxy is horrible, and it's tacked onto tor in the most frankenstein-y fashion possible
22:14:40  <dg> I don't like their pluggable transports but the tech is cool
22:14:48  <dg> (emulating Skype is one cool thing)
22:17:27  <psi> iirc obsproxy can be counter productive
22:17:33  <hottuna> i've gotta go
22:17:46  <psi> due to it emulating a survalence network
22:17:53  * psi spelling
22:18:00  <dg> bye
22:19:55  <psi> i've got to part for now as well
22:20:27  <dg> I think we can just call it a day now and have it next week/sometime soon(er)
22:20:34  <dg> Kind of fucked this one up
22:21:04  <iRelay> <weltende@freenode> sounds like a plan
22:21:07  <psi> it's best to have a predefine meeting structure
22:21:54  <dg> yeah
22:26:10  <Meeh> disconnected...
22:26:29  <Meeh> 23:10:30 <+psi> in general a generic transport api would be a developer's goldmine
22:26:32  <Meeh> 23:10:31 < hottuna> we have a lot of random data, but none of the headers of ssl
22:26:35  <Meeh> 23:13:01 <+Meeh> yea, yet.. it's a matter of time
22:26:38  <Meeh> 23:13:15 <+Meeh> so why make people offline from i2p for a while.. better safe than sorry
22:26:41  <Meeh> what did I miss?
22:27:11  <psi> Meeh: meeting adjurned for now
22:27:11  * psi spelling
22:27:58  <psi> <dg> certainly not, lillith
22:27:58  <Meeh> ah, lame.. meeting next week?
22:28:10  * psi lag
22:28:55  <lillith> Meeh, to be decided, maybe before since this one wasn't a great success
22:29:25  <Meeh> true true, next week then