351 lines
32 KiB
HTML
351 lines
32 KiB
HTML
{% extends "_layout.html" %}
|
|
{% block title %}I2P Development Meeting 156{% endblock %}
|
|
{% block content %}<h3>I2P dev meeting, November 15, 2005</h3>
|
|
<div class="irclog">
|
|
<p>15:15 < jrandom> 0) hi</p>
|
|
<p>15:15 < jrandom> 1) Net status / 0.6.1.5</p>
|
|
<p>15:15 < jrandom> 2) Syndie updates</p>
|
|
<p>15:15 < jrandom> 3) I2Phex</p>
|
|
<p>15:15 < jrandom> 4) I2P-Rufus</p>
|
|
<p>15:15 < jrandom> 5) Issue tracker</p>
|
|
<p>15:15 < jrandom> 6) Dynamic Keys</p>
|
|
<p>15:15 < jrandom> 7) ???</p>
|
|
<p>15:15 < jrandom> 0) hi</p>
|
|
<p>15:15 * jrandom waves</p>
|
|
<p>15:16 < jrandom> weekly status notes posted up @ http://dev.i2p.net/pipermail/i2p/2005-November/001210.html</p>
|
|
<p>15:17 <+bar> yalla! *fires some rounds into the air*</p>
|
|
<p>15:17 * jrandom ducks and covers, diving into 1) Net status / 0.6.1.5</p>
|
|
<p>15:18 < jrandom> as mentioned in the mail, there's been a lot of progress, and there should be a new release later tonight</p>
|
|
<p>15:18 * jrandom would have released it earlier, but I slept late and didn't want everyone upgrading /during/ the meeting :)</p>
|
|
<p>15:20 < jrandom> anyone have any questions/comments/concerns re: 1) net status / 0.6.1.5?</p>
|
|
<p>15:20 <+fox> <ailouros> is "please keep up the good work" an acceptable comment?</p>
|
|
<p>15:20 < jrandom> :) thanks</p>
|
|
<p>15:22 < jrandom> I've been pretty happy with the stability as of late. hopefully the next release will improve throughput beyond 4-8KBps/stream. I've done plenty of local testing, but we need to see it out in the wild</p>
|
|
<p>15:22 < tethra> i second ailouros's comment, and furthermore, propose a toast:</p>
|
|
<p>15:22 < jrandom> we've also had some more positive reports from users on dialup connections</p>
|
|
<p>15:22 < tethra> to jrandom, and i2p! woot!</p>
|
|
<p>15:22 < tethra> <3</p>
|
|
<p>15:23 < jrandom> w3wt. ok, if there's nothing else, lets jump on over to 2) Syndie updates</p>
|
|
<p>15:24 < jrandom> lots of progress on this front, but perhaps it'll be best to discuss it after the release when people can try it for themselves</p>
|
|
<p>15:25 < jrandom> hopefully the info up @ http://syndiemedia.i2p.net/about.html (the first link) can explain why you should bother trying it out :)</p>
|
|
<p>15:25 <+fox> <ailouros> oh come on, first you don't release it, then you say "try it first"... this is just teasing! :D</p>
|
|
<p>15:25 < jrandom> :)</p>
|
|
<p>15:26 < jrandom> ok ok, so lets just jump ahead to 3) I2Phex then, so y'all can post up your thoughts about syndie to syndie itself after you upgrade ;)</p>
|
|
<p>15:27 < jrandom> there's going to be an announcement for I2Phex 0.1.1.36 later tonight</p>
|
|
<p>15:28 < jrandom> the only change is the fix for the annoying "Please insert a disk" popup</p>
|
|
<p>15:28 < tethra> that means i can take the disk out the drive without it screaming at me, then? ;)</p>
|
|
<p>15:28 < jrandom> heh yes</p>
|
|
<p>15:28 < tethra> :D</p>
|
|
<p>15:30 < jrandom> ok, if there's nothing more on 3) I2Phex, lets jump on over to 4) I2P-Rufus</p>
|
|
<p>15:30 < tethra> what are the plans for i2phex, while we're on the subject?</p>
|
|
<p>15:30 < jrandom> ah</p>
|
|
<p>15:30 < jrandom> there's a set of feature requests posted to the forum</p>
|
|
<p>15:31 < jrandom> I haven't heard anything from redzara about the code merge with Phex, but Gregor is still working on abstracting the networking stuff so we can more easily keep in sync</p>
|
|
<p>15:32 < jrandom> generally, the app seems functional, though gwebcache support would be Really Good, so that I2Phex could work out of the box without needing to fetch any files or keys</p>
|
|
<p>15:32 < jrandom> I don't know anyone working on getting gwebcache support (back) into I2Phex, but if someone knows java, that'd be Really Useful</p>
|
|
<p>15:33 < tethra> cool.</p>
|
|
<p>15:33 <+fox> <reliver> _007pig perhaps ?</p>
|
|
<p>15:33 <+fox> <ailouros> sorry if I ask, but wasn't gnutella network the one that flooded itself to death some time ago?</p>
|
|
<p>15:33 < tethra> the new guys do tend to be a bit confused about it at first</p>
|
|
<p>15:33 <+fox> <reliver> you did not take him up on his offer for help, yesterday, jrandom</p>
|
|
<p>15:33 < jrandom> _007pig was looking into translation work, but anyone would be great. Phex itself has gwebcache support, but sirup disabled it</p>
|
|
<p>15:34 < jrandom> ailouros: gnutella is still around, but yeah, its not ideal.</p>
|
|
<p>15:34 < tethra> is anyone looking into perhaps changing the protocol i2phex uses to something else?</p>
|
|
<p>15:35 < jrandom> I'm hesitant to demand people work on specific projects, so I instead suggest a few different areas that someone could explore</p>
|
|
<p>15:35 < jrandom> tethra: no one that I know of</p>
|
|
<p>15:35 <+fox> <ailouros> well, I think I'd rather see Localhost (azureus modification) on i2p then</p>
|
|
<p>15:36 < tethra> surely bittorrent is more awkward than gnutella?</p>
|
|
<p>15:36 < tethra> in terms of seeding and such</p>
|
|
<p>15:36 < jrandom> ailouros: whatever people implement and maintain is good :)</p>
|
|
<p>15:36 <+fox> <ailouros> I don't know, I didn't use gnutella since... 6 years I think</p>
|
|
<p>15:37 < anti> surely it is more efficient and better test of true scalability?</p>
|
|
<p>15:37 <+fox> <ailouros> jrandom yeah that's a good metric :D</p>
|
|
<p>15:37 < jrandom> i2phex works pretty well, I've transferred lots of data through it, and found some neat content</p>
|
|
<p>15:37 <@cervantes> (pony pr0n)</p>
|
|
<p>15:37 <+fox> <ailouros> lol</p>
|
|
<p>15:37 < tethra> hahah</p>
|
|
<p>15:37 < jrandom> there may be better ways to do things, but something that works is better than something that doesn't exist</p>
|
|
<p>15:37 < tethra> cervantes++</p>
|
|
<p>15:37 < tethra> ;)</p>
|
|
<p>15:38 < tethra> truer words have never been spoken.</p>
|
|
<p>15:39 < anti> good point</p>
|
|
<p>15:39 <@cervantes> uhoh... jr has taken offense and gone early to dinner</p>
|
|
<p>15:39 <@cervantes> (sorry)</p>
|
|
<p>15:39 < anti> no, he's probably searching for that (mythical) pony pr0n. ;)</p>
|
|
<p>15:40 < jrandom> *cough* ;)</p>
|
|
<p>15:40 < tethra> lol </p>
|
|
<p>15:40 < tethra> heheh ;)</p>
|
|
<p>15:40 < jrandom> ok, if there's nothing else on 3), lets move on to 4) I2P-Rufus</p>
|
|
<p>15:40 <+fox> <reliver> i want flying pony pr0n :-)</p>
|
|
<p>15:40 < jrandom> Rawn / defnax: anything to add to what was posted on the forum?</p>
|
|
<p>15:41 <@cervantes> looks like some good progress is being made</p>
|
|
<p>15:41 < jrandom> aye</p>
|
|
<p>15:45 < jrandom> ok, if there's nothing on that, lets jump on to 5) issue tracker</p>
|
|
<p>15:45 < jrandom> the forum is a bit heavyweight for managing bugs and feature requests, and bugzilla is a bit of a beast... </p>
|
|
<p>15:46 <@frosk> isn't there a bugzilla already somewhere?</p>
|
|
<p>15:46 < jrandom> i've posted up some general requirements, and cervantes has come up with one workable solution</p>
|
|
<p>15:46 < jrandom> nah, the bugzilla was on the old host (@johnscompanies) before we migrated to sago</p>
|
|
<p>15:46 <+fox> <ailouros> hot about NNTP? better than forums, usually threaded...</p>
|
|
<p>15:46 <+fox> <reliver> strange that bugzilla is so lacking, considering the huge open source community using it ...</p>
|
|
<p>15:46 <+fox> <ailouros> how*</p>
|
|
<p>15:46 <@frosk> ah ok</p>
|
|
<p>15:47 < jrandom> nntp has potential, but there are some benefits over that by using syndie (simple filtering by tag): http://syndiemedia.i2p.net:8000/threads.jsp?visible=ovpBy2mpO1CQ7deYhQ1cDGAwI6pQzLbWOm1Sdd0W06c=/1132012800004&post=ovpBy2mpO1CQ7deYhQ1cDGAwI6pQzLbWOm1Sdd0W06c=/1132012800004&</p>
|
|
<p>15:48 < jrandom> but nntp does have the benefits of having decades of battle testing</p>
|
|
<p>15:48 <+fox> <ailouros> NNTP reader filter by keyword (the [] tags)? :D</p>
|
|
<p>15:49 <@modulus> perhaps not so much testing of late?</p>
|
|
<p>15:49 <+fox> <reliver> including spamming and flaming ...</p>
|
|
<p>15:49 < jrandom> we'd want something web accessible though, since most people don't use nntp readers</p>
|
|
<p>15:49 <+fox> <ailouros> I say Thunderbird is good in that sense, and you can share the enigmail between i2mail and i2nntp</p>
|
|
<p>15:49 <@modulus> maybe a web accessible nntp reader?</p>
|
|
<p>15:49 <+fox> <reliver> gateways are common</p>
|
|
<p>15:49 < jrandom> hmm modulus?</p>
|
|
<p>15:50 <@modulus> well, usenet is not so much used anymore i think</p>
|
|
<p>15:50 < jrandom> right, so we'd have to have an nntp server and a gateway with filtering support</p>
|
|
<p>15:50 <@frosk> i like cervantes' idea though</p>
|
|
<p>15:50 <+fox> <ailouros> (and I also say the reason people don't use NNTP readers is because forums are so much prettier and so much heavier)</p>
|
|
<p>15:50 <@modulus> hmm, gateway with filtering support? what are you guys talking about, maybe it helps knowing. :-)</p>
|
|
<p>15:51 <@modulus> imo forums suck, i hate fucking forums, they're unusable ;-(</p>
|
|
<p>15:51 <+fox> <ailouros> LOL I guess he wants the access from the InterNEt</p>
|
|
<p>15:51 <+fox> * ailouros agrees with modulus</p>
|
|
<p>15:51 <@frosk> modulus: so very true</p>
|
|
<p>15:51 < jrandom> heh modulus ;) we're discussing http://syndiemedia.i2p.net:8000/threads.jsp?visible=ovpBy2mpO1CQ7deYhQ1cDGAwI6pQzLbWOm1Sdd0W06c=/1132012800004&post=ovpBy2mpO1CQ7deYhQ1cDGAwI6pQzLbWOm1Sdd0W06c=/1132012800003&</p>
|
|
<p>15:51 <+fox> <ailouros> aieee the megabyte long URI</p>
|
|
<p>15:52 <@modulus> what I love about syndie URLs is how memorable and simple they are to type</p>
|
|
<p>15:52 < jrandom> I do still like http://syndiemedia.i2p.net:8000/threads.jsp?post=ovpBy2mpO1CQ7deYhQ1cDGAwI6pQzLbWOm1Sdd0W06c=/1132012800004&</p>
|
|
<p>15:52 < jrandom> heh</p>
|
|
<p>15:52 < jrandom> well, go to http://syndiemedia.i2p.net/threads.jsp then and click on the "Issue tracking software" link :)</p>
|
|
<p>15:53 <@frosk> bug reporting right from your router console</p>
|
|
<p>15:53 <@modulus> hmm, bug tracking.</p>
|
|
<p>15:53 < jrandom> using syndie would give us 1) integration with every I2P user's environment 2) trivial filtering 3) threading 4) spam handling (via ignore/favorites) 5) syndie a workout :)</p>
|
|
<p>15:54 <+fox> <reliver> sounds great :-)</p>
|
|
<p>15:54 <+fox> <ailouros> it is</p>
|
|
<p>15:54 < jrandom> aye that is a really good feature frosk... we could even have specialized html forms to post to /syndie/post.jsp</p>
|
|
<p>15:54 <+fox> <ailouros> and by the way, wasn't there talk about basing syndie on NNTP? :D :D :D</p>
|
|
<p>15:54 <@modulus> hmm, how about the Debian bug tools? they're nice i think, the mailbug</p>
|
|
<p>15:54 < anti-> can't argue with what already works!</p>
|
|
<p>15:55 <@cervantes> I think you should do it purely from a techdemo perspective</p>
|
|
<p>15:55 < jrandom> ailouros: using NNTP to distribute syndie posts, yeah. right now we just use ad-hoc syndication, but further enhancements would be great</p>
|
|
<p>15:56 <@cervantes> no better way to demonstrate syndie than with some real world use cases</p>
|
|
<p>15:56 < jrandom> true enough</p>
|
|
<p>15:56 < jrandom> ok, perhaps we can plan on getting that out in the 0.6.1.6 release</p>
|
|
<p>15:56 <+fox> <reliver> what i don't like about forum is they are low entry cost</p>
|
|
<p>15:57 <+fox> <reliver> so lots of distractions filling them.</p>
|
|
<p>15:57 <@modulus> i don't know, this syndie thing ... i much do not like yet, but maybe i'll get used to it.</p>
|
|
<p>15:57 <+fox> <reliver> and you can only work with them online</p>
|
|
<p>15:57 < jrandom> modulus: have you read the post linked to from http://syndiemedia.i2p.net/about.html ?</p>
|
|
<p>15:57 <@modulus> reliver: high-entry is bad for bug reports though, people are making you a big favour by bothering to report in a sense.</p>
|
|
<p>15:57 <+fox> <ailouros> they are not low entry cost: bandwidth comes to mind. They are high noise levels, so you can use [font=54]HELLO WORLD![/font] and annoy a huge number of people in no time</p>
|
|
<p>15:57 < jrandom> agreed modulus</p>
|
|
<p>15:58 <+fox> <ailouros> oh yeah and you have to be online indeed</p>
|
|
<p>15:58 < jrandom> heh ailouros, thats something we need to deal with in Syndie anyway :)</p>
|
|
<p>15:58 <@modulus> hmm, probably not, jr, let me check</p>
|
|
<p>15:58 <+fox> <ailouros> well, with syndie you can blacklist the users and you're pretty much set</p>
|
|
<p>15:58 < jrandom> well, with syndie you can create your bug reports offline, then syndicate them up to a remote archive later when you are :)</p>
|
|
<p>15:58 < jrandom> exactly ailouros, with one click in the new release too</p>
|
|
<p>15:59 <+fox> <ailouros> with forums either you hope for an admin to come and kill'em, or you keep them</p>
|
|
<p>15:59 < anti-> it's more uucp than nntp :)</p>
|
|
<p>15:59 <@modulus> hmm, which post in particular linked from there?</p>
|
|
<p>15:59 < jrandom> lol *exactly* anti</p>
|
|
<p>15:59 < jrandom> modulus: the first link "in syndie itself"</p>
|
|
<p>15:59 * cervantes likes the killing option</p>
|
|
<p>16:00 <@modulus> bah, uucp == nntp for all practical purposes :-)</p>
|
|
<p>16:00 < jrandom> anti-: thats actually the point - as people build newer and better transport mechanisms (uucp, nntp, usenetdht, etc), the content can flow seamlessly</p>
|
|
<p>16:00 <+fox> <ailouros> this all reminds me of plan9</p>
|
|
<p>16:01 <+fox> <reliver> i2p may be special, but usually bug reporting systems used as firewalls against users ...</p>
|
|
<p>16:01 < jrandom> used as firewalls against users?</p>
|
|
<p>16:01 <+fox> <reliver> i2p may be special, but usually bug reporting systems are used as firewalls against users ...</p>
|
|
<p>16:01 <+fox> <reliver> yes.</p>
|
|
<p>16:01 < jrandom> I want it to be really, really easy for people to report bugs</p>
|
|
<p>16:01 <+fox> <reliver> mozilla, thunderbird, ubuntu are just examples</p>
|
|
<p>16:02 <+fox> <reliver> ok, great :-)</p>
|
|
<p>16:02 < jrandom> mozilla/etc have that integrated "feedback agent" for submitting bug reports automatically</p>
|
|
<p>16:02 <+fox> <reliver> they don't read those bug reports</p>
|
|
<p>16:02 < jrandom> heh</p>
|
|
<p>16:02 <@modulus> hmm, that intro is ok, only problem is i just don't like the interface at all, i prefer doing mailish things through the folder metaphor rather than the web-with-sithloads-of-links-on-it method</p>
|
|
<p>16:02 <@modulus> but that's just me</p>
|
|
<p>16:02 < jrandom> modulus: perhaps the rss export would best serve your needs then?</p>
|
|
<p>16:02 <+fox> <ailouros> I agree with modulus (anyone guessed? :D )</p>
|
|
<p>16:02 <@cervantes> having to use pastebin to show console errors is a bit of a put-off for some folks</p>
|
|
<p>16:03 < jrandom> or we can get susimail integration, as cervantes suggested, to send out reports</p>
|
|
<p>16:03 < jrandom> (or to post to syndie)</p>
|
|
<p>16:03 <@modulus> it is possible, jrandom, i'll look into it. maybe i need an RSS-to-NNTP or RSS-to-POP?/IMAP converter, i'll think on it.</p>
|
|
<p>16:05 <@cervantes> modulus: I'll be curious to find out what you think of the new i2ptunnel interface come the next i2p release</p>
|
|
<p>16:05 <@cervantes> whether it's better or worse for you in terms of usability</p>
|
|
<p>16:05 <@cervantes> (but I guess you just normally edit the config files?)</p>
|
|
<p>16:07 < jrandom> ooh yeah shit, I forgot so much stuff in the status notes...</p>
|
|
<p>16:08 <+fox> <ailouros> then let's hurry ahead and skip to the next point in line... that was point number C, right?</p>
|
|
<p>16:08 * jrandom thinks it really kicks ass, but we'll get some more feedback as people try it out</p>
|
|
<p>16:08 <@modulus> cervantes: is that curious as in "you're going to kill yourself with a small knife in your arse as a better alternative to using it" or on the contrary? :-)</p>
|
|
<p>16:08 < jrandom> yeah, jumping to 6), anyone have any thoughts on the Dynamic Keys proposal?</p>
|
|
<p>16:09 <@modulus> cervantes: usually use the interface actually, though now i know the config files are editable ... :-)</p>
|
|
<p>16:09 <+fox> <ailouros> yeah, I'm pretty certain it will cause the skyrocket in the number of supposed known routers</p>
|
|
<p>16:09 <@cervantes> *damn* :)</p>
|
|
<p>16:10 <@modulus> this dynamic key is the idea that routers get a new key upon new IP, right?</p>
|
|
<p>16:10 <@cervantes> modulus: well, just if it's even worth bothering with WAI bullshit</p>
|
|
<p>16:10 < jrandom> heh thats true ailouros</p>
|
|
<p>16:10 <@cervantes> anyway...I digress</p>
|
|
<p>16:10 < jrandom> right modulus </p>
|
|
<p>16:11 <@modulus> well, perhaps it isn't bad that the known peers are actually guesswork, more so than now.</p>
|
|
<p>16:11 <+Complication> Well, the only thing I can figure out about Dynamic Keys.. seems that one shouldn't change keys needlessly (or it screws reliability performance tracking).</p>
|
|
<p>16:11 <+Complication> But when IP changes (rare enough?) it might not hurt.</p>
|
|
<p>16:11 < jrandom> right Complication. it isn't something we'd want by default. most people will *not* want it</p>
|
|
<p>16:12 < anti-> i'm not sure of the positive impact of the proposals.</p>
|
|
<p>16:12 < jrandom> it won't offer much of an improvement for anonymity either, and no improvement at all against a powerful adversary, but it might help against weak adversaries</p>
|
|
<p>16:12 <+fox> <ailouros> wouldn't it also give away which nodes are fixed ip and which aren't?</p>
|
|
<p>16:13 * cervantes has had the same key for nearly 2 years :)</p>
|
|
<p>16:13 <+polecat> Well at least I can get here.</p>
|
|
<p>16:13 < jrandom> ailouros: it would not be used by most people. only a very, very small minority would want to use it</p>
|
|
<p>16:13 <+fox> <ailouros> so basically more churn for a bit of protection against weak adversaries?</p>
|
|
<p>16:13 < jrandom> right ailouros</p>
|
|
<p>16:13 <+fox> <ailouros> oh ok</p>
|
|
<p>16:14 <+fox> <ailouros> is there a way to measure the performance hit of that feature once in the wild?</p>
|
|
<p>16:14 <@modulus> it would, i think, help against a node-dest intersection attack?</p>
|
|
<p>16:14 <+polecat> I still wonder why I keep switching between OK and OK(NAT), puzzling...</p>
|
|
<p>16:14 < jrandom> modulus: only for a weak adversary</p>
|
|
<p>16:14 <+fox> <ailouros> polecat don't worry, I keep switching between 15h uptime and 0h uptime :|</p>
|
|
<p>16:14 < jrandom> ailouros: not sure, though stats.i2p suggests that we can handle the churn</p>
|
|
<p>16:15 < jrandom> polecat: hmm, means there's likely some filtering going on</p>
|
|
<p>16:15 <@modulus> imo the node-dest intersection attack is the most serious likely feasible attack atm? besides the fact we are too few, i mean.</p>
|
|
<p>16:15 <@modulus> so, i think anything which helps on that line is probably a good idea</p>
|
|
<p>16:16 <+polecat> I can send UDP packets right over my router at that port, no problem from remote shells. No clue, perhaps i2p detects the NAT, and mistakenly thinks it isn't forwarded.</p>
|
|
<p>16:16 <+fox> <ailouros> I agree with the "good idea" as long as the churn doesn't cause a severe performance hit</p>
|
|
<p>16:16 < anti-> when the network is bigger, there will be plenty of churn anyway...</p>
|
|
<p>16:17 < anti-> *points out the obvious DoS attack involving constantly changing keys every few minutes</p>
|
|
<p>16:17 < anti-> what impact would that have?</p>
|
|
<p>16:17 <+fox> <ailouros> dos against who? :D</p>
|
|
<p>16:18 < jrandom> eh, new peers go in the "not failing" tier by default, and only go up to the "high capacity" or "fast" tiers after they are around for a while</p>
|
|
<p>16:18 < jrandom> so it won't DoS peer selection</p>
|
|
<p>16:18 < anti-> with a relatively strong opponent... would create an awful lot of apparently dead nodes/netdb churn?</p>
|
|
<p>16:18 <+Complication> anti: nobody would consider that node reliable any more</p>
|
|
<p>16:18 <+polecat> anti-: We have a shitlist for a reason.</p>
|
|
<p>16:19 < anti-> *satisfied</p>
|
|
<p>16:19 < jrandom> well, the netDb entries are dropped if the peer is unreachable</p>
|
|
<p>16:20 < anti-> then the same performance issues that were just raised about dynamic keys would apply? if the performance wouldn't be too impacted by such an attack, the performance wouldn't be affected noticeably by dynamic keys either... would it?</p>
|
|
<p>16:20 <+polecat> incremental trust really does help with handling late onset betrayers, I was thinking.</p>
|
|
<p>16:20 <+fox> <ailouros> what's a "late onset betrayer"?</p>
|
|
<p>16:20 <+polecat> Trust people more and more as they continue to benefit you, but never so much that they can take away more than they've given...</p>
|
|
<p>16:20 < anti-> join for ages, then turn judas.</p>
|
|
<p>16:21 < jrandom> right, peers get dropped out of the 'fast' tier quickly if they act poorly</p>
|
|
<p>16:21 <+Complication> I'd think it would be someone behaving like "wait until 300 participating tunnels, crash"</p>
|
|
<p>16:21 <+polecat> Oh, I make up phrases all the time. Yeah, Judas type betrayal, where you genuinely help someone, then betray them with the idea of cashing in at the last minute.</p>
|
|
<p>16:21 < anti-> oh no, the tunnels broken *rebuild*</p>
|
|
<p>16:21 < jrandom> the peers promoted to the 'fast' tier during that time they're dropped should then suffice</p>
|
|
<p>16:21 <+fox> * ailouros has fun with these incorrect bible refernces :D</p>
|
|
<p>16:22 < jmg> speaking of high capacity, wow im getting between 400k and 600K constantly for the router today. (but maybe all those zero hops settings im using are helping)</p>
|
|
<p>16:22 < jrandom> 600KBps?!</p>
|
|
<p>16:22 <+polecat> Hopefully during the time it takes to get to 300 participating tunnels, you'll be required to help transfer enough data it wouldn't matter if you crashed.</p>
|
|
<p>16:22 < jmg> yes</p>
|
|
<p>16:22 <+fox> <ailouros> O_O what are you connected to?</p>
|
|
<p>16:22 <+Complication> Such bandwidth is news to me :)</p>
|
|
<p>16:22 < jrandom> damn, thats fast enough to start running into our bloom filters</p>
|
|
<p>16:22 < anti-> ailouros: rude question to anony researchers ;)</p>
|
|
<p>16:23 <+polecat> It's gotta be 600KBpm or ph.</p>
|
|
<p>16:23 <+fox> <ailouros> sorry anti- :D but he was the first to speak</p>
|
|
<p>16:23 <+polecat> puh!</p>
|
|
<p>16:23 < jrandom> I'd love to get some stats from the oldstats.jsp page off you. but glad to hear its handling things :)</p>
|
|
<p>16:23 < anti-> one day i will try from i2...</p>
|
|
<p>16:23 < jrandom> hehe</p>
|
|
<p>16:24 <+fox> <ailouros> sounds cool, I2P on I2</p>
|
|
<p>16:24 < jmg> jrandom: im keeping graphs, ill monitor more closely, but yes i can confirm 600kB/s sustained for 2 minutes, about 5 minutes ago</p>
|
|
<p>16:24 <+polecat> Has anyone tried to traverse a d-link router's firewall? I'm having no luck there whatsoever and my friend keeps forgetting to forward the port.</p>
|
|
<p>16:24 < jrandom> nice jmg </p>
|
|
<p>16:24 < anti-> polecat: do we do udp holepunching yet? i lost track</p>
|
|
<p>16:25 < jrandom> anti-: yes, we do, for all but symmetric NATs</p>
|
|
<p>16:25 < jrandom> polecat: if your friend has their model #, there are a few sites online listing what type of NAT it is</p>
|
|
<p>16:26 < anti-> regarding late onset betrayal... might be an issue with a powerful adversary?</p>
|
|
<p>16:26 < jmg> jrandom: of course bittorrent has been known to rape this connection at 4MB/s sustained, but Iv eased up on that a little lately</p>
|
|
<p>16:26 < anti-> 24000 nodes, so you get one crashing every 10 seconds or so?</p>
|
|
<p>16:26 <+polecat> symmetric NAT, as opposed to full cone?</p>
|
|
<p>16:26 < jrandom> nice jmg </p>
|
|
<p>16:26 < jrandom> hmm anti-?</p>
|
|
<p>16:26 < jrandom> polecat: or restricted cone</p>
|
|
<p>16:27 <+polecat> Wow, it can even do restricted cone that's impressive..</p>
|
|
<p>16:27 < anti-> i don't think late onset betrayal would have any significant effect at all unless applied on an incredibly massive scale, at which other attacks would have more of an impact?</p>
|
|
<p>16:28 < jrandom> yeah I'm not too worried about it anti-... it'd cost too much, and we can route around failures anyway, so the damage would be minimal</p>
|
|
<p>16:28 <+Complication> Late betrayal kind of requires contributing a lot (as to get other machines relying on your machine).</p>
|
|
<p>16:28 <+fox> <ailouros> incredibly massive scale = you are all the netries on almost everyone else's router?</p>
|
|
<p>16:28 < anti-> that is exactly what anti-p2ps do now, but we do have anti-anti-p2ps now...</p>
|
|
<p>16:29 <+fox> <ailouros> no wait anti-p2p send trash instead of good data</p>
|
|
<p>16:29 <+fox> <ailouros> that's not the same</p>
|
|
<p>16:29 < anti-> that's just a faster way of getting shitlisted, so you would never be listed well.</p>
|
|
<p>16:29 < anti-> that wouldn't work against i2p at all, i think.</p>
|
|
<p>16:29 <@cervantes> jmg: I've had 4-5mb/s off torrents before, but never anything like 600k over I2P...have you got beefy hardware too?</p>
|
|
<p>16:29 <+polecat> I was more thinking independant of i2p persay. My government does a lot of late onset betrayal, though they try to keep it classified.</p>
|
|
<p>16:29 < anti-> but we would probably bleed them dry of bandwidth first!</p>
|
|
<p>16:29 < jrandom> anti-: if they're reliable for days on end, they can only attack once for less than 10 minutes</p>
|
|
<p>16:30 < jrandom> exactly anti- :)</p>
|
|
<p>16:30 <+polecat> Or in the context of online banking.</p>
|
|
<p>16:30 < jmg> does anyone have easy instructions on setting up the Native BigInteger library for amd64? if not ill just figure it out</p>
|
|
<p>16:30 < jrandom> heh polecat </p>
|
|
<p>16:30 < jrandom> jmg: its built into jbigi.jar, but it should build on amd64 now</p>
|
|
<p>16:30 < jrandom> though, I suppose this means we're now on 6.1) ??? </p>
|
|
<p>16:31 < jrandom> anyone have anything else to bring up? :)</p>
|
|
<p>16:31 < anti-> you'd need 20000 machines or something, with a rolling crash schedule, and i think the results would be disappointing; you would end up contributing far more to the network than you took away!</p>
|
|
<p>16:31 < jrandom> that is the hope anti-</p>
|
|
<p>16:31 <+fox> <ailouros> well, worst case scenario is that people must reseed</p>
|
|
<p>16:31 < jmg> oh thanks</p>
|
|
<p>16:31 <+polecat> 64 bit processor, 4mbit upload bandwidth, sounds like somebody's a lucky bastard.</p>
|
|
<p>16:32 < anti-> or running a normal machine at a uni...</p>
|
|
<p>16:32 <+fox> * ailouros looks at his uni's hardware list and frowns</p>
|
|
<p>16:32 < anti-> a uni that doesn't buy dell ;)</p>
|
|
<p>16:33 <+fox> <ailouros> I think we have a couple of dells... from 5 years ago IIRC</p>
|
|
<p>16:33 <+fox> <Sonium> i think this is bad:</p>
|
|
<p>16:33 <+fox> <Sonium> jvm 1 | java.lang.OutOfMemoryError</p>
|
|
<p>16:33 <+fox> <Sonium> jvm 1 | java.lang.OutOfMemoryError</p>
|
|
<p>16:33 <+fox> <Sonium> jvm 1 | java.lang.OutOfMemoryError</p>
|
|
<p>16:33 <@cervantes> polecat: 4 megabyte ;-)</p>
|
|
<p>16:33 < jrandom> Sonium: yeah, once it gets one OOM, it'll die fast</p>
|
|
<p>16:34 <+fox> <Sonium> and this too:</p>
|
|
<p>16:34 <+fox> <Sonium> jvm 1 | 21:21:44.484 CRIT [ Establisher] sport.udp.EstablishmentManager: Err</p>
|
|
<p>16:34 <+fox> <Sonium> or in the establisher</p>
|
|
<p>16:34 < jrandom> (subsequent OOMs are safe to ignore)</p>
|
|
<p>16:34 < jrandom> once it gets a single OOM, you can ignore all subsequent errors</p>
|
|
<p>16:34 <+fox> <ailouros> yeah but you shouldn't have the first OOM :D</p>
|
|
<p>16:34 < jmg> polecat: the latency out here on the russian space station in phenominal though..</p>
|
|
<p>16:34 < jrandom> true ailouros</p>
|
|
<p>16:35 <+fox> <ailouros> oh, by the way... my router gets watchdogged quite often</p>
|
|
<p>16:35 < jrandom> hrm, high cpu usage?</p>
|
|
<p>16:35 <+fox> <ailouros> I guess it's just my unlucky installation?</p>
|
|
<p>16:35 <+fox> <ailouros> not that I know of, the machine is rather unloaded</p>
|
|
<p>16:36 <+fox> <ailouros> but I guess this is what I should expect from a buggy JVM on a somewhat bugged linux emulation layer</p>
|
|
<p>16:36 < jrandom> what jvm are you using, and what os?</p>
|
|
<p>16:36 <+fox> <Sonium> me?</p>
|
|
<p>16:36 <+fox> <ailouros> Sun's Java(tm) 2 Standard Edition, JRE 5.0 Update 5 on NetBSD/i386 2.0.2</p>
|
|
<p>16:37 < jrandom> ahhh yeah, I have done no testing on nbsd. fbsd is fine, but I don't have any experience w/ nbsd</p>
|
|
<p>16:38 < jrandom> might be worth trying out gcj, perhaps we can dig into that after the meeting</p>
|
|
<p>16:38 <+fox> <ailouros> it works rather well, but the real fun with this is that sometimes (depending on which bit he flipped when getting off the bed -- err restarting) the netbsd files get created with 540 permission :D</p>
|
|
<p>16:38 <+fox> <Sonium> something really sucks here</p>
|
|
<p>16:38 <+fox> <Sonium> jvm 1 | # Internal Error (53414645504F494E540E4350500175), pid=3500, tid=345</p>
|
|
<p>16:38 <+fox> <Sonium> 6</p>
|
|
<p>16:39 <+fox> <ailouros> sorry the netDb files are created 540</p>
|
|
<p>16:39 <+fox> <Sonium> I think I will reinstall this later</p>
|
|
<p>16:39 < jrandom> Sonium: what OS are you on? the jvm seems to be acting up</p>
|
|
<p>16:39 <+fox> <Sonium> winxp</p>
|
|
<p>16:39 < jrandom> yeah, if you're on 1.5.0_5, might be worth trying 1.4.2_09</p>
|
|
<p>16:39 < anti-> i don't think that's i2p's problem...</p>
|
|
<p>16:40 < jrandom> (1.4.2 has been more stable for me, requiring less resources)</p>
|
|
<p>16:40 < jrandom> and i2p doesn't use any 1.5-isms, nor do we need the 1.5 GUI improvements</p>
|
|
<p>16:40 <+fox> <Sonium> the curious thing is, that is never occured before</p>
|
|
<p>16:40 <+polecat> Can't use azureus if you don't have 1.5 though, meh.</p>
|
|
<p>16:40 <+fox> <ailouros> and of course I *DO* use azureus :|</p>
|
|
<p>16:41 <+fox> <ailouros> but it isn't a real problem... not much, I think...</p>
|
|
<p>16:41 <+fox> <ailouros> unless those messages about bob being fourth are relevant</p>
|
|
<p>16:41 < jrandom> nah, those are safe to ignore</p>
|
|
<p>16:41 < anti-> (am i the only one irked by utorrent and bitcomet not being open?)</p>
|
|
<p>16:42 <+polecat> :o Damn you bob!</p>
|
|
<p>16:42 < jrandom> ok, anyone have anything else for the meeting?</p>
|
|
<p>16:42 < anti-> muffins?</p>
|
|
<p>16:42 * cervantes can recommend ibm java 1.4.2 if you're after better resource handling</p>
|
|
<p>16:42 <+polecat> anti-: Try mlnet. caml -> weirdest language in the world, but it works well.</p>
|
|
<p>16:42 <+fox> <ailouros> caml is cool</p>
|
|
<p>16:42 <+fox> <ailouros> (if you can read it :D )</p>
|
|
<p>16:42 <@frosk> hey, don't diss caml</p>
|
|
<p>16:43 < anti-> prolog deserves a mention there, as does brainf**k et al</p>
|
|
<p>16:43 <+polecat> caml has horrible docs. It took me half an hour to figure out that ! usually (sometimes) is a dereference operator.</p>
|
|
<p>16:43 <@frosk> i'm paid to write ocaml :)</p>
|
|
<p>16:43 <+polecat> jrandom: Didn't know I crashed a meeting, sorry.</p>
|
|
<p>16:44 < jrandom> np, we're making up for our short meetings ;)</p>
|
|
<p>16:44 * jrandom winds up</p>
|
|
<p>16:44 * jrandom *baf*s the meeting closed</p>
|
|
</div>
|
|
{% endblock %} |