I2P dev meeting, February 8, 2005

Quick recap

  • Present:

ant, brachtus, cervantes, duck, jrandom, jrandom2p, legion, modulus, postman, protokol, Ragnarok, smeghead, ugha2p,

完全な IRC ログ

13:05 < jrandom> 0) hi
13:05 < jrandom> 1)*
13:05 < jrandom> 2) 0.5
13:05 < jrandom> 3) i2p-bt 0.1.6
13:05 < jrandom> 4) fortuna
13:05 < jrandom> 5) ???
13:06 < jrandom> 0) hi
13:06  * jrandom waves
13:06 <@duck> y0
13:06 < smeghead> hi
13:06 < jrandom> weekly status notes up @ http://dev.i2p.net/pipermail/i2p/2005-February/000564.html
13:07 < cervantes> sorry I'm late...I was busy reading the status notes that were posted at the last minute...
13:07 < jrandom> hey, this week they were /before/ the meeting at least (by 30s or so ;)
13:08 < jrandom> anyway, while you dig through that oh so exciting email, lets jump on into 1)*
13:09 < jrandom> with the latest patches from anon et al, i'm torn between pushing out a new so close to the 0.5 rev.  
13:10 < jrandom> for the moment though, if you're feeling brave, feel free to give cvs a whirl - its stable (i'm breaking things off on another branch), and has some good stuff
13:11 < jrandom> the deciding factor for not pushing a rev out was when i did a checklist for 0.5 and found that the only things left were really web interface updates
13:11 <+Ragnarok> about the patches from sugadude, they do represent a policy change, as we discussed filtering out non .i2p addresses before, and you decided against it
13:11 < jrandom> oh, hrm?  i disagree with my old self then - eepproxy doesn't accept non-.i2p address in any case, even if they were in hosts.txt
13:12 < jrandom> did i have a convincing argument before?
13:13 <+Ragnarok> ok, then can we revert the patch, and I can implement it the way it originally worked, which is a 0 line change?
13:13 <+Ragnarok> not really, I just didn't care either way :)
13:13 < jrandom> oh, cool you're the boss
13:13 < cervantes> well you convinced me to drop all my work on a multi-tld management system and fire all my employees
13:13 <+Ragnarok> filtering is already happening, so it's just adding a condition to an if statement
13:14 < jrandom> cervantes: there's also this beautiful bridge i've got for sale...
13:14 < cervantes> :)
13:14 < jrandom> ok word Ragnarok, if you want to send me a .java/.tar/.diff/.whatever, that'd be great
13:15 <+Ragnarok> I can do cvs now :)
13:15 < jrandom> :) even better
13:15  * cervantes backs up cvs head
13:15 < jrandom> heh
13:16 <+Ragnarok> *BOOM*
13:16 <+Ragnarok> ... just kidding :)
13:17 < jrandom> ok, other than that, anyone have anything else to bring up wrt 0.4.*?
13:17 < ant> <dm>gt; 0.4.* sucks, give us 0.5
13:17 < ant> <dm>gt; It's like a gazillion years old!!
13:18 < ant> <fvw>gt; 0.4.* doesn't suck, give us 0.5 anyway.
13:18 < jrandom> 2) 0.5 it is then :)
13:19 < ant> <dm>gt; you guys owe me big time, I brought 0.5
13:19 < jrandom> we couldn't'a done it without ya dm 
13:19 < ant> <dm>gt; amen
13:20 < jrandom> as mentioned in the notes, pretty much all the heavy lifting for 0.5 is done and tested, but there are still the odds and ends left to fix up
13:21 < jrandom> (e.g. the next task on my list is a tunnel config page to manage the pools and settings)
13:22 <@duck> I hope we will have a test-0.5 network before releasing?
13:22 < jrandom> there have been updates to lots of different components though, so 0.5 might be a bit bumpy
13:22 < ant> <dm>gt; jrandom HAS a test network already.. duh
13:23 < jrandom> aye, i've been doing one locally here with a dozen routers, but in the next day or two i'll try to snag some people to help with some wide area tests
13:24  * postman can offer a dedicated machine
13:24 < jrandom> wikked.  perhaps we can try something out tomorrow, try to break some things.  
13:26 < cervantes> as can I
13:27 < jrandom> word
13:27 < jrandom> thats about all i have to say about the upcoming 0.5 at the moment - the cvs commit logs have been pretty verbose, so if you want the nitty gritty, hit 'em up
13:28 < jrandom> anyone else have any comments/questions/concerns/frisbees wrt 0.5?
13:29 <+postman> no
13:29  * postman is looking forward to get the new V8 running :)
13:30 < jrandom2p> well, 0.5 is more of a new tank - designed to improve security and anonymity, not as a performance tweak ;)
13:30 < jrandom2p> but i agree, its been too long
13:30 <@duck> dont forget to add a 0.5 target to bugzilla
13:30 <@duck> in case there are bugs
13:30 < jrandom2p> (heh, did i even add a 0.4?)
13:31 < jrandom2p> but good call
13:31 <@duck> or would you like bugs elsewhere
13:31 <@duck> err bugreports :)
13:31 <@duck> I know that I have been lazy and abuse irc messages for them
13:31 < jrandom2p> no, bugzilla is great, much better than my notebook
13:32 < jrandom2p> i don't blame you, as bugzilla is a bit of a pain
13:32 < jrandom2p> but as bugs pile up, its for the best
13:32 <@duck> nah
13:33  * jrandom just noticed i'm switching schitzophrenically between screens
13:34 < jrandom> ok, anyway, moving on to 3) i2p-bt 0.1.6
13:34 < jrandom> duck: you've got the mic 
13:34 <@duck> ok
13:34 <@duck> i2p-bt 0.1.5 had some issues, the two biggest ones:
13:35 <@duck> - resource temporarily unavailable
13:35 <@duck> - invalid argument error on windows
13:35 <@duck> both have been fixed
13:35 < jrandom> (yay!)
13:35 <@duck> while I tried to blame the sam protocol, the sam bridge and winsock
13:35 <@duck> the problem turned out to be related to non-blocking socket code
13:36 <@duck> I yet have to see 0.1.6 crash
13:36 <@duck> some other issues are not addressed:
13:36 <@duck> the GUI users have been complaining about the popups
13:36 <@duck> you can comment them out, but I didnt like that
13:37 <@duck> still waiting for someone to implement a better solution
13:37 <@duck> like showing a status line on the transfer window itself
13:37  * smeghead hides
13:37 < smeghead> i looked at that last night actually
13:37 < smeghead> but it's not at the top of my priority list
13:37 <@duck> or maybe one day I will look into how wxPython works and do it myself
13:37 <@duck> but it's not at the top of my priority list
13:38 <@duck> and I dont use the GUI, so I dont really care :P
13:38 <+Ragnarok> there's always the new gui from 3.9 :)
13:38 <@duck> is it any better?
13:38 < smeghead> yes why did you base i2p bt on such a crusty version in the first place? :)
13:38 <@duck> because it was the stable release at that moment
13:39 <@duck> and not as mutilated as clients like bittornado
13:40 <@duck> Ragnarok: ignoring licensing issues, I think that porting our i2p things to 3.9 might be good
13:40 <+Ragnarok> the new gui is pretty awsome, imho, and it's written using pygtk, so I can actually hack on it
13:40 < jrandom> what's 3.9's license?  i thought it was mit-esque?
13:40 <+protokol> i would love a more recent jetty version
13:40 < smeghead> protokol: that's coming sooner than you think
13:41 <@duck> "BitTorrent Open Source License"
13:41 < smeghead> flavor of the month license
13:41 <+Ragnarok> I haven't read all of it.. it seems odd
13:41 <+protokol> licencing does not exist on i2p
13:41 <@duck> derived from the Jabber Open Source License 1.0
13:41 <+protokol> if there is source, its PD
13:41 <@duck> protokol: that is why I said 'ignoring'
13:42 < smeghead> and the jabber license is based on?
13:42 < jrandom> (out of date copyright laws?)
13:42 < smeghead> besides that :)
13:43 < modulus> Sun's wish to fuck about.
13:43 <@duck> http://www.opensource.org/licenses/jabberpl.php
13:43 < smeghead> i move we schedule the licensing issue for the next meeting of the I2P Public Domain Security Council
13:43 < modulus> ah, that one
13:43 < modulus> misheard.
13:45 <@duck> 3.9.0 looks hot
13:45 <@duck> it is still beta though
13:47 <@duck> ok, those willing to help, please let me know
13:47 <@duck> so we can look into using 3.9.x
13:47 <@duck> .
13:47 < jrandom> w3rd
13:47 < smeghead> i'm willing to help out
13:47 < jrandom> i'm willing to help test
13:48 <+Ragnarok> I'm willing, but there are likely to be time constraints, as I am currently having the semester from hell.
13:48 < jrandom> d'oh
13:48 <@duck> drop out
13:48 < jrandom> damn, duck beat me
13:48 < smeghead> yes, everyone does it
13:49 <+Ragnarok> boo
13:49 < ant> <jnymo>gt; just join the military ;)
13:50 < jrandom> yeah, as that'll give you lots of time to code, 'eh?  ;)
13:50 <+Ragnarok> I've already given up on being a math major, that's as much as you're getting from me :)
13:50 < jrandom> heh
13:50 < jrandom> ok, anyone else have anything on 3) i2p-bt?  
13:51 < ant> <jnymo>gt; just don't sign up for six years
13:51 <@duck> quite a bit of forum posts on it
13:51 <@duck> thanks to those who aid the newbies
13:51 <@duck> s/thanks/my thanks/
13:51 <@duck> if you have stuff for a FAQ, lemme kno
13:52 < jrandom> (if we still had drupal, we could just add a new node...)
13:53 < jrandom> ok, anyway, moving on to 4) fortuna
13:54 < jrandom> smeghead: wanna give us an update on things?
13:54 < smeghead> yes, i'm working on pants and fortuna in tandem
13:55 < smeghead> since i needed to modify fortuna's build to turn it into a pbuild
13:55 < smeghead> eta on a patch that will let you test fortuna is a day or two, maybe tonight depending on what drugs are involved
13:56 < jrandom> heh
13:56 <@duck> so you'll get your pants down?
13:56 < jrandom> ok, cool, whenever is fine - if we get it in for 0.5 in the next week or so, thats great, if not, thats great too
13:56 < smeghead> well even if i finish it tonight, i would take a conservative stance on deployment
13:57 < jrandom> reasonable enough
13:57 < smeghead> until we get some decent testing in
13:57 < smeghead> since this will be at the heart of most of i2p's crypto
13:57 < jrandom> aye
13:57 < ant> <jnymo>gt; will jbigi stay?
13:57 < smeghead> your new entropy class is cool
13:58 < jrandom> yeah jnymo, this is just a random # generator
13:58 < ant> <jnymo>gt; ah
13:59 < jrandom> we'll still need to do some research into the quality of various entropy sources in the router, but I think we'll be able to feed it some data.
14:00 < smeghead> btw if anyone wants to read what this pants thing is about: http://smeghead.i2p/README_pants
14:00 < jrandom> oh wikked
14:01 < smeghead> pants is almost done too
14:01 < brachtus> i know jbigi is kinda hard to get working with OS X/Darwin... will this have the same build problems?
14:01 < smeghead> what is the issue on osx?
14:01 < modulus> it's just you have to build the lib
14:02 < modulus> not a big deal imo, but somewhat troublesome.
14:02 < jrandom> brachtus: fortuna is in pure java, doesnt use anything native
14:02 < smeghead> i can put jbigi into pants and that should make building a cinch if we ship pants with i2p
14:02 < brachtus> nothign terribly difficult, it's like building a shared lib on linux, but harder than just double-click-install
14:02 < smeghead> you'd need ant of course
14:02 < brachtus> ok jrandom, that's great :)
14:03 < jrandom> smeghead: thats actually a good point - jbigi has a pants dependency upon GMP
14:03 < ant> <jnymo>gt; what is pants?
14:03 < smeghead> no manual mucking would be necessary
14:03 < ant> * jnymo doesn't have a router up
14:03 < smeghead> jnymo: read that link i just posted
14:04 < jrandom> http://bolas.mine.nu:8080/cgi-bin/nph-proxy/000000A/http/smeghead.i2p/README_pants
14:04 < smeghead> pants can build gmp too
14:04 < jrandom> (public inproxy)
14:04 < smeghead> ah nice
14:04 < jrandom> yuck, that totally b0rked the text
14:04 < ant> <jnymo>gt; thanks jr
14:04 < ant> <fvw>gt; aren't you afraid of legal trouble?
14:04 < smeghead> jrandom doesn't run the inproxy
14:04 < jrandom> oh, the inproy is run by someone else, its been posted to the forum
14:05 < jrandom> (see http://bolas.mine.nu:8080/)
14:05 < cervantes> jrandom: it shouldn't be viewed as an html file...check the source
14:05 < ant> <fvw>gt; still, I'm amazed anyone would. But as long as it's being run by someone not vital to the project, fine :)
14:05 < jrandom> hehe
14:05 < jrandom> we're /all/ vital to the project :)
14:06 < smeghead> fvw: i don't see inproxies as legally precarious as outrpoxies
14:06 < smeghead> outproxies even
14:06 < ant> <fvw>gt; Perhaps not, but they can still serve up child porn and such
14:06 < jrandom> only if there were such things on i2p, which, to my knowledge, there isnt 
14:06 < legion> outproxies could route through tor, just to be a little safer, since they would just be used for webrowsing I don't see it as a problem.
14:07 < jrandom> (but yeah)
14:07 < modulus> yet
14:07 < ant> <fvw>gt; yeah, but anyone can put it on at any point.
14:07 < ant> <fvw>gt; yeah, I wouldn't run a tor outproxy either. Anyway, sorry for drifting offtopic like that
14:07 < jrandom> legion: yeah, though i tossed up squid.i2p before tor was out
14:07 < ant> <duck_>gt; to get back on topic; looking forward to pants
14:08 < jrandom> aye, pants++
14:08 < smeghead> i'll let you know before i drop pants on CVS
14:08 < smeghead> it's kinda big
14:08 < ant> <duck_>gt; folks outside of i2p might be interested in it too
14:09 < cervantes> yes let us all know before you drop your pants
14:09 < smeghead> yes, i intend to publicise it outside of i2p also
14:09 < jrandom> agreed, perhaps we should put it in another module (or on the new fast/large server)?
14:09 <+Ragnarok> especially if you're a big pants kind of guy
14:10 < smeghead> yes the pants module really should be kept separate from the pants repo in the source tree, currently i have them located in the same apps/pants root
14:10 < smeghead> :/
14:10 < smeghead> which i don't have to tell you is total pants
14:11 < smeghead> so what were we talking about originally?
14:11 < jrandom> hmm, we can discuss deployment options offline
14:11 < jrandom> fortuna ;)
14:11 < smeghead> right
14:12 < jrandom> smeghead: have you looked at the AES/SHA256 needs of the impl?
14:12 < jrandom> (as i2p's SHA256 doesn't do partial digests)
14:13 < smeghead> hm
14:13 < jrandom> AES we've got perfectly suitable block impl though
14:13 < smeghead> i guess i'll find out when it blows up
14:13 < jrandom> anyway, we can work those through too
14:13 < jrandom> heh
14:15 < jrandom> ok, anyone have any questions/thoughts/concerns on fortuna?
14:15 < jrandom> if not, hopping on over to 5) ???
14:15 < jrandom> cervantes: p1ng
14:16 < cervantes> http://forum.i2p/viewtopic.php?t=305
14:16 < cervantes> we have a new forum member of the week
14:16 < cervantes> I present [drumroll] Sugadude!
14:16  * brachtus applauds Sugadude
14:17 < jrandom> yay
14:17 < cervantes> for generally being a helpful sod to all those i2p n00bs
14:17 <@duck> nice avatar too
14:17 < cervantes> avatar(s)
14:18 < legion> avatars? didn't know that we could have avatars on the i2p forums?
14:18 < smeghead> only users who are really really bad get them
14:18 < cervantes> you can't...unless you're a forum person of the week ;-)
14:18 <@duck> only for the elite
14:18 < legion> oh, i see...
14:19 < ant> <jnymo>gt; i know someone was interested in secure financial systems over i2p
14:19 < legion> makes sense :)
14:19 < ant> <jnymo>gt; don't know if they're here, but...
14:19 <@duck> I am a smelly anarcho capitalist
14:19 <@duck> so try me
14:20 < ant> <jnymo>gt; i was reading more on threashold cryptography and theres talk about using it for that
14:20 < ant> <jnymo>gt; as well as securing other functions
14:21 < ant> <jnymo>gt; everyone familiar with threshold cryptography?
14:21 < legion> IMO that cryptography and network security should be variable, how much should depend on the feature/task.
14:21 < ant> <duck_>gt; jnymo: a bit
14:22 < ant> <jnymo>gt; well, for trustable financial transactions in i2p, we want strong decentralized trust
14:22 < modulus> is that about the shared keys and shit like that?
14:23 < ant> <jnymo>gt; yea, keys are shared in pieces
14:23 < ant> <duck_>gt; but in an anonymous environment, how do you know that the entities doing the sharing arent controlled by the same one?
14:23 < ant> <jnymo>gt; and you need to circumvent more than half of all the servers in the system to obtain the priv key
14:24 < modulus> afaik it's kind of complicated the issue of distributed key generations though.
14:24 < legion> yeah but in a system of millions that would be hard (yeah i2p is small at the moment, but hopefully it will grow much larger soon).
14:25 < ant> <jnymo>gt; atomic communications, or something.. but yea, theres issues with taking on new nodes on the system, which i thing are being worked out
14:25 < ant> <jnymo>gt; think
14:25 < ant> <jnymo>gt; so maybe its not developed enough, but i'd bet some usage of threshold crypto will end up over i2p at some point
14:26 < jrandom> neat
14:26 < legion> dunno, maybe
14:26 < ant> <jnymo>gt; someone has already built a DNSSEC addon with threshold crypto
14:27 < ant> <jnymo>gt; and a wrapper around bind
14:27 < jrandom> thresholds work fine when identity is scarce
14:27 < jrandom> in anonymous networks, however, identity is free
14:27 < legion> I'd figure at the moment the highest priority is to get it more user friendly and debugged.
14:27 < jrandom> (want a new destination?  want 100,000?)
14:28 < legion> granted it's cool whenever a new service/feature is developed.
14:28 < jrandom> aye, commerce and finance on top of i2p will be nice
14:28 < ant> <jnymo>gt; yea, and i wouldn't know if atomic commo would work over a 10000 node threshold crypto sys
14:29 < ant> <jnymo>gt; well, that's all i had to say :)
14:30 < jrandom> heh cool, definitely feel free to post up neat stuff to the forum or whatnot whenever
14:30 < jrandom> ok, anyone else have anything for the meeting?
14:32 <+ugha2p> I suck.
14:33 < jrandom> whats up ugha2p?
14:33 < ant> <jnymo>gt; glad you got that off your' chest, ugha ;)
14:33 <+ugha2p> I never remember the meetings. :)
14:33 < jrandom> heh
14:33 < jrandom> well, the logs will be posted soon, 90 minutes of action packed fun
14:34 < jrandom> well, on that note
14:34  * jrandom winds up
14:34  * Curiosity waves to jrandom and stays thank-you! :D
14:34 < ant> * jnymo pitches the meeting ball
14:34  * jrandom *baf*s the meeting closed