I2P dev meeting, June 8, 2004 @ 21:02 UTC

Quick recap

  • Present:

cervantes, deer, duck, fvw, hypercubus, mihi, Nightblade, Sonium, ugha_node,

Jurnal IRC complet

21:02:08 <duck> Tue Jun  8 21:02:08 UTC 2004
21:02:21 <duck> meeting time
21:02:33 <duck> writeup is at http://dev.i2p.net/pipermail/i2p/2004-June/000268.html
21:02:39 <duck> but I did make a mistake in the numbering
21:02:45 <duck> so the first item 5 will be skipped
21:02:53 <hypercubus> yay!
21:03:03  * duck puts some ice in his beer
21:03:14  * mihi 'd rename first #5 to #4 ;)
21:03:27 <hypercubus> nah, let's just have two item 4's next week ;-)
21:03:37  * duck renames 'hypercubus' to 'mihi'
21:03:48 <hypercubus> yay!
21:03:49 <duck> ok
21:03:53 <duck> * 1) libsam
21:04:02 <duck> is there a Nightblade in the channel?
21:04:39 <duck> (idle     : 0 days 0 hours 0 mins 58 secs)
21:05:03 <hypercubus> ;-)
21:05:53  * duck reclaims the microphone
21:06:15 <duck> Nightblade wrote a SAM lib for C / C++
21:06:23 <duck> it compiles for me.. but that is all I can say :)
21:06:37 <mihi> no test cases? ;)
21:07:06 <duck> if there are any rFfreebsd users Nightblade might be interested in you
21:07:08 <ugha_node> The strstr calls really annoyed me in the code. ;)
21:07:27 <ugha_node> duck: What's a rFfreebsd?
21:07:42 <duck> how I did type freebsd
21:08:00 <mihi> rm -rF freebsd?
21:08:29 <ugha_node> Too bad -F doesn't work with rm.
21:08:30 <duck> ugha_node: its bsd licensed; so fix it
21:08:41 <fvw> sounds sane to me :). Alas I uninstalled my last freebsd box a while back. I 
                have accounts on other peoples' boxes though, and am willing to run testcases.
21:08:43 <ugha_node> duck: I might. :)
21:08:50 <duck> (damn BSD hippies)
21:09:09 <duck> oh, nice and short frank
21:09:17 <duck> mo libsam comments?
21:09:49 <duck> fvw: I guess Nightblade will contact you if he has a need
21:09:50  * fvw grumbles at perfectly sane unix behaviour for killing his irc client.
21:10:02 <duck> but since his email was a week old be might have found something
21:10:17 <mihi> fvw: ?
21:10:24 <fvw> yeah, if someone wanted to take me up on my offer I sort of missed that. Feel 
                free to send email or something.
21:10:42  * duck hops to #2
21:10:46 <hypercubus> uhm, to where? ;-)
21:10:54 <duck> 2) browse i2p and normal web with one browser
21:10:57 <fvw> fresh install, haven't yet told my zsh not to hup stuff in the bacgrkground. 

21:11:09 <fvw> hypercubus: I'm on the public mailinglist user list I think. fvw.i2p@var.cx
21:12:11 <duck> there was some stuff about adding all TLDs to your brower proxy ignore list
21:12:23 <fvw> does that require discussion? I think it was pretty much handled on the 
21:12:24 <duck> I think it is a dirty hack
21:12:36 <fvw> yes, that was mentioned. Welcome back.
21:12:47 <duck> fvw: I didnt read the thread :)
21:13:12 <duck> okay, if you dont want to discuss it, move to #3
21:13:19 <duck> * 3) chat channel
21:13:23 <hypercubus> cervantes' script works perfectly on Konqueror 3.2.2, Firefox 0.8, and 
                       Opera 7.51, all for Gentoo w/KDE 3.2.2
21:13:39  * mihi places a flag on #4
21:13:55 <duck> #i2p-chat is an alternative channel here for offtopic chat and light support
21:14:08 <duck> I dont know who did regg it
21:14:12 <hypercubus> i did
21:14:17 <duck> so better be careful :)
21:14:22 <fvw> ehm, there is no #4, just two #5's :)
21:14:33 <hypercubus> i'll be lucky if i can remember the password when i need it ;-)
21:14:33 <mihi> [22:27] -ChanServ-      Channel: #i2p-chat
21:14:33 <mihi> [22:27] -ChanServ-      Contact: hypercubus <<ONLINE >>

21:14:33 <mihi> [22:27] -ChanServ-    Alternate: cervantes <<ONLINE >>
21:14:37 <mihi> [22:27] -ChanServ-   Registered: 4 days (0h 2m 41s) ago
21:15:12 <hypercubus> i gave a few trusted peeps op powers for whenver i'm not around and 
                       there's trouble
21:15:24 <duck> sounds good
21:15:39 <duck> it might be a bit overkill
21:15:51 <hypercubus> you never know on IRC ;-)
21:15:55 <duck> but after this protogirl did join here I thought it would be good to clean up 
                 this chan
21:16:03 <hypercubus> heh
21:16:27 <hypercubus> we'll need it for sure sometime in the next few months anyhow
21:16:34 <duck> jups
21:16:48 <duck> and then the freenode ppl will kick us out 
21:16:55 <hypercubus> ;-)
21:17:13 <duck> they dont like anything that isnt written in their kampf
21:17:16 <duck> err
21:17:44  * duck moves to $nextitem and triggers mihi's breakpoint
21:17:47 <hypercubus> i figured tying the new channel in with support would legitimize it for 
21:18:47 <duck> hypercubus: you might be surprised
21:19:04 <hypercubus> *cough* i admittedly didn't read all the policies...
21:19:24 <duck> it is russian roullete
21:19:39 <hypercubus> hmm, didn't think it would be quite that dire
21:19:52  * duck is being negative
21:19:54 <hypercubus> well i'll look into what we can do
21:20:09 <fvw> sorry, I must have missed something. Why would freenode kick us off?
21:20:21  * duck looks at the timeout counter for mihi's breakpoint
21:20:32 <duck> fvw: they focus on development channels
21:20:35 <mihi> ?
21:20:53 <mihi> duck: the breakpoint triggers on /^4).*/
21:21:01 <duck> mihi: but there is no #4
21:21:06 <fvw> so? i2p is soo alpha that right now even support is development.
21:21:11 <fvw> (and no, you may not quote me on that)
21:21:36 <duck> fvw: you might not be familiar with the types of discussion that did happen 
                 on IIP
21:21:38 <hypercubus> yeah but we have *2* channels for it
21:21:45 <duck> and which will likely happen in #i2p channels
21:22:04 <duck> I am pretty sure that freenode does not appreciate it.
21:22:10 <Nightblade> i'm here now
21:22:49 <hypercubus> we'll donate a margarita machine to them or something
21:22:49 <mihi> duck: what do you refer to? the floods? or #cl? or what?
21:23:08 <fvw> discussions on IIP or discussions on #iip? I've never seen anything apart from 
                devel and support on #iip. And discussions on IIP would move to I2P, not 
21:23:09 <duck> all kinds of non political correct talk
21:23:36 <fvw> there's margarita machines? Ooh, me want.
21:23:54 <duck> oh well
21:24:38 <hypercubus> shall we revisit 2)?
21:24:58 <duck> hypercubus: what do you have to add about the browser proxy?
21:25:18 <hypercubus> oops, number 1... since nightblade just graced us with his presence ;-)
21:25:33 <duck> Nightblade: we took the freedom to 'discuss' libsam
21:25:42 <Nightblade> Ok, i'll say a few lines
21:25:48 <hypercubus> but yeah i had something that wasn't brought up on the list about the 
                       browser thing too now that i think about it
21:25:56 <duck> Nightblade: fvw told us that he might be able to help with some freebsd 
21:26:20 <fvw> I don't have a freebsd machine anymore but I have accounts on freebsd 
                machines, give me test cases and I'd be happy to run them.
21:27:02 <Nightblade> I have started working on a C++ dht, which uses Libsam (C).  At this 
                       point I have not gotten especially far although I've been working a lot 
                       on it.  right now nodes in the dht can "ping" each other through a sam 
                       data message
21:27:09 <Nightblade> in the process i found a couple minor bugs in libsam
21:27:18 <Nightblade> which i will post a new version of sometime in the future
21:27:51 <ugha_node> Nightblade: Could you please remove those 'strstr' calls from libsam? :)
21:27:52 <Nightblade> the test case is: try to compile it and report the errors to me
21:28:01 <Nightblade> what is wrong with strstr
21:28:21 <ugha_node> It's not meant to be used instead of strcmp.
21:28:38 <Nightblade> oh yeah, also I am going to port libsam to windows, but that is not in 
                       any near future
21:29:07 <Nightblade> is there anything wrong with the way i am using it, besides aesthetics?
21:29:15 <Nightblade> you can send me changes or tell me what you'd rather do
21:29:19 <Nightblade> that just seemed the easiest way
21:29:21 <ugha_node> Nightblade: I didn't notice any.
21:29:32 <fvw> strcmp is more efficient than strstr ofcourse.
21:29:36 <ugha_node> But I just skimmed through it.
21:30:20 <ugha_node> fvw: You can occasionally exploit stuff which uses strstr instead of 
                      strcmp, but that's not the case.
21:31:22 <Nightblade> yeah now i see some places where i can change it
21:31:28 <fvw> that too, but I'm assuming you'd have noted that. Well, actually, you'd have 
                to use strncmp to prevent those exploits. But that's besides the point.
21:31:31 <Nightblade> i don't remember why i did it that way
21:31:57 <ugha_node> fvw: I agree.
21:32:27 <Nightblade> oh now i remember why
21:32:40 <Nightblade> it is a lazy way of not having to figure the length for strncmp
21:32:49 <duck> heh
21:32:52 <ugha_node> Nightblade: Heheh.
21:33:01 <fvw> use min(strlen(foo), sizeof(*foo))
21:33:04 <hypercubus> shall the spanking commence?
21:33:15 <fvw> I thought the oral sex came first? *ducks*
21:33:32 <fvw> right, next point I think. Hypercube had a comment about proxying?
21:33:38 <hypercubus> heh
21:33:54 <duck> bring it on!
21:34:03 <Nightblade> i will make the changes for the next version - change some of them at 
21:34:25 <hypercubus> ok, well this had been discussed briefly in channel a few weeks back, 
                       but i think it bears revisiting
21:34:48 <deer> * Sugadude volunteers to perform the oral sex.
21:34:59 <hypercubus> rather than adding TLD's to your browser's block list, or using the 
                       proxy script, there's a third way
21:35:29 <hypercubus> which shouldn't have the same drawbacks as the other two approaches 
21:36:17 <fvw> which I'll tell you for the cheap cheap price of $29.99? Spill it already!
21:36:27 <hypercubus> and that would be to have the eeproxy re-write incoming html pages to 
                       embed the page in a frameset... 
21:36:58 <hypercubus> the main frame would contain the requested HTTP content, the other 
                       frame would serve as a control bar
21:37:13 <hypercubus> and would allow you to turn on/off proxying at will
21:37:40 <hypercubus> and will also alert you, perhaps via colored borders or some other kind 
                       of alert, that you're browsing non-anonymously
21:37:54 <fvw> how are you going to prevent an i2p site (with javascript etc) from turning 
                off anonimity?
21:37:59  * duck tries to apply jrandom-skill-level-of tolerance
21:37:59 <hypercubus> or that a link in an eepsite page leads to the RealWeb(tm)
21:38:04 <duck> cool! make it!
21:38:16 <fvw> you'll still have to do something fproxy-like, or make something 
                non-browser-controlled for switching.
21:38:29 <ugha_node> fvw: Right.
21:39:10 <hypercubus> that's why i'm throwing this out here again, perhaps someone might have
                        some ideas about how to secure this
21:39:31 <hypercubus> but imo this is something that will be sorely needed for most i2p end 
21:39:33 <hypercubus> *users
21:40:04 <hypercubus> because the TLD/proxy script/dedicated browser approaches are too much 
                       to ask of your general net user
21:40:29 <fvw> In the long run, I think an fproxy workalike is the best idea. But that's 
                definately not a priority imho, and I don't actually think browsing sites will 
                be the i2p killer app.
21:40:42 <Sonium> What is the netDb anyway?
21:40:59 <duck> Sonium: database of known routers
21:41:10 <hypercubus> fproxy is too cumbersome for most users
21:41:32 <Sonium> doesn't such a database compromise annonymity?
21:41:39 <hypercubus> imo it's part of the reason freenet never caught on in the non-dev 
21:41:41 <fvw> hypercube: not necessarily. proxy autoconfiguragion ("pac") can make it as 
                simple as filling in a single value in your browser config. I think we 
                shouldn't underestimate the fact that in the foreseeable future, all the i2p 
                users will be at least slightly clueful computer-wise. (all evidence on 
                freenet-support notwithstanding)
21:42:00 <ugha_node> Sonium: No, 'bad guys' could collect that information manually anyway.
21:42:21 <Sonium> but if NetDb is down i2p is down, right?
21:42:29 <fvw> hypercubus: Not really, I think the fact that it hasn't worked at all since 
                early 0.5 is more to blame for that. </offtopic time="once again">

21:42:44 <fvw> Sonium: you can have more than one netdb (anyone can run one)
21:42:58 <hypercubus> we already have pac, and even though it works spectacularly from a tech 
                       standpoint, realistically it's not going to protect the anonymity of 
                       the avg. jog
21:43:03 <hypercubus> *avg. joe
21:43:22 <ugha_node> fvw: Err.. Every router has its own netDb.
21:43:42 <duck> ok. I am about to pass out. be sure to *baff* the meeting closed after you 
                 are done
21:43:52 <ugha_node> I2P has no central dependencies anymore.
21:44:07 <hypercubus> ok, well i just wanted to get this idea formally in the logs ;-)
21:44:30 <fvw> ugha_node: ok, a published netdb then. I don't actually run a node (yet), I'm 
                not entirely up with the terminology.
21:44:34 <ugha_node> Hmm. Didn't mihi want to say something?
21:45:05  * fvw feeds duck coffee-flavoured chocolate to keep him up and running a little bit 
21:45:07 <mihi> no :)
21:45:21 <mihi> is duck a network device? ;)
21:45:25 <ugha_node> mihi: Btw, are you going to take the window size increase bounty?
21:45:28  * fvw feeds duck alcohol-flavoured chocolate to shut him down indefinately.
21:45:30 <hypercubus> in swedish
21:45:52 <mihi> ugha_node: what bounty?
21:46:00 <hypercubus> okay, then on to 5), rant-a-rama? ;-)
21:46:13 <ugha_node> mihi: http://www.i2p.net/node/view/224
21:46:27  * duck eats some of fvw's chocolate
21:47:16 <mihi> ugha_node: definitely no; sorry
21:47:36 <ugha_node> mihi: Uh, okay. :(
21:48:33  * mihi tried to hack up the "old" streaming api some time ago, but that one was too 
21:48:53 <mihi> but it would imho be easier to fix that one instead of fixing mine...
21:49:21 <ugha_node> Heh.
21:49:42 <hypercubus> so modest
21:49:46 <mihi> since it already has some (broken) "reordering" support in it
21:50:49 <Sonium> is there a way to ask deer how many people are on the i2p-#i2p channel?
21:51:01 <duck> no
21:51:08 <hypercubus> nope, but i can add that to bogobot
21:51:08 <Sonium> :/
21:51:11 <Nightblade> !list
21:51:13 <deer> <duck> 10 ppl
21:51:13 <hypercubus> after i finish the installer ;-)
21:51:24 <Sonium> !list
21:51:32 <Sonium> o_O
21:51:35 <mihi> Sonium ;)
21:51:38 <ugha_node> This is not an fserv channel!
21:51:39 <Sonium> that was a trick!
21:51:40 <ugha_node> :)
21:51:41 <hypercubus> should be !who
21:51:44 <deer> <duck> ant duck identiguy Pseudonym ugha2p bogobot hirvox jrandom Sugadude 
21:51:48 <cervantes> oop missed the meeting
21:51:57 <ugha_node> !list
21:52:01 <Nightblade> !who
21:52:11 <deer> <duck> !who-your-mom
21:52:17 <mihi> !who !has !the !list ?
21:52:21 <fvw> !yesletsallspamthechannelwithinoperativecommands
21:52:33 <Nightblade> !ban fvw!*@*
21:52:42 <mihi> !ban *!*@*
21:52:50 <hypercubus> i sense a gavel coming down
21:52:51 <duck> sounds like a good time to close it down
21:52:55 <Sonium> btw, you should also implement an !8 command like chanserv has
21:52:59 <fvw> right, now we have that settled, let's clo.. yes. that.
21:53:00  * hypercubus is psychic
21:53:05 <duck> *BAFF*
21:53:11 <Nightblade> !baff
21:53:12 <hypercubus> my hair, my hair
21:53:24  * fvw points at hypercube and laughs. Your hair! Your hair!