I2P dev meeting, December 7, 2004

Quick recap

  • Present:

ant, bushka, clayboy, dinoman, duck, Frooze, mule, postman, protokol, Ragnarok, slart, ugha2p,

IRC 完整日志

22:00:00 <@duck> Tue Dec  7 21:00:00 UTC 2004
22:00:04 <@duck> I2P meeting time
22:00:05 < Frooze> i just made Frooze up for i2p.  i don't even know what a 'frooze' is.
22:00:21 <@duck> as announced on http://dev.i2p.net/pipermail/i2p/2004-December/000509.html
22:00:29 <@duck> Agenda:
22:00:29 <@duck> 0) hi
22:00:29 <@duck> 1)
22:00:29 <@duck> 2) i2p-bt
22:00:29 <@duck> 3) #idlerpg
22:00:29 <@duck> 4) ???
22:00:32 <@duck> .
22:01:09 <@duck> 0) hi
22:01:15 < clayboy> hi
22:01:16 <@duck> jrandom called in sick
22:01:20 <+ugha2p> Hi.
22:01:30 <@duck> plus msged me that he'd probably not make it
22:01:39 <+protokol> http://www.google.com/search?q=frooze
22:01:41 <@duck> so we'll see and just start
22:01:46 < clayboy> hope he gets better quick
22:02:06 <@duck> 1)
22:02:16 <@duck> new release will be out Real Soon
22:02:31 <@duck> so tomorrow or thursday.
22:02:41 <@duck> there has been quite a few bugfixes
22:03:24 <+ugha2p> Do newer CVS revisions also fix the memory/CPU issues?
22:03:29 < clayboy> a few of us have been following the cvs builds, it's working very nicely
22:03:33 <@duck> most streaming lib, sam bridge, etc
22:04:17 <+ugha2p> I've been experiencing some uncommon loads from I2P.
22:04:23 < clayboy> i think those were fixed many revisions ago, ugha2p
22:04:41 <+ugha2p> (Running -7)
22:04:51 < clayboy> oh, hm
22:04:52 <@duck> ugha2p: dont see anything about that in the history
22:05:48 <+protokol> you know what would be nice (if not feasable/worth it) is an RSS feed of the changelog
22:05:48 <@duck> ok
22:05:49 <+ugha2p> That's strange.
22:06:01 <+protokol> ;-)
22:06:17 <@duck> maybe file a bugzilla item
22:06:25 <@duck> or dunno
22:06:34 <+ugha2p> The Java process consumes 100% of CPU for about half of the time.
22:07:18 <+ugha2p> So, you don't know anything about the issue? Do your routers behave OK?
22:07:24 < dinoman> yea it is high for me to -6
22:08:24 <@duck> top/uptime info is behaving weird for me since my nptl upgrade, so cant say
22:09:03 <+ugha2p> Ok, maybe we should move on?
22:09:07 <@duck> ok
22:09:14 <@duck> 2) i2p-bt
22:09:24 <+ugha2p> And ask jrandom when he is about to release
22:09:40 <+ugha2p> It has worked fine for me with NPTL.
22:09:45 <@duck> ugha2p: he said tomorrow or thursday
22:09:58 <+ugha2p> Right.
22:09:59 <@duck> yesterday I released a new i2p-bt
22:10:23 <@duck> I gained some new understanding of the whole 'buffer' concept
22:10:42 <@duck> plus there were some previous pending patches from Ragnarok
22:11:13 < mule> duck: congratulations, good work!
22:11:15 <@duck> also the slice size is increased, which means that instead of sending 32KB each time, it sends 128KB
22:11:29 <@duck> which should keep the queue filled
22:11:47 <+ugha2p> Yeah, thanks, duck. :)
22:11:56 <@duck> DrWoo and others filed some GUI feature requests
22:12:23 <@duck> but I never use the GUI myself, wouldnt know wxpython and probably dont care too much :)
22:12:31 <+Ragnarok> fitting each slice into a single message didn't work as well as expected?
22:12:57 < clayboy> many seeded torrents on http://brittanyworld.i2p/bittorrent/ if anyone want to try (with i2p and i2p-bt 0.1.3)
22:13:10 <@duck> Ragnarok: it is a bit of a guess
22:13:27 <@duck> it gives much higher throughput values on local transfers
22:13:51 <+ugha2p> Maybe we should wait for someone to port a full-featured client instead?
22:14:10 <+Ragnarok> hm, ok
22:14:13 <@duck> we can all wait :)
22:14:37 < clayboy> BitTorrent _is_ "full featured", it's the only client i use for bt (also off i2p) :)
22:15:15 <+ugha2p> clayboy: Not really. :)
22:16:02 <@duck> personally I prefer things with sound defaults
22:16:17 <@duck> take mldonkey, you can change 1 million things and most users have no idea what they do
22:16:50 <@duck> this leads to user-myths, like i2p users hitting 'Reseed' all the time, or reinstalling if it doesnt work
22:17:01 <+ugha2p> If you aren't willing to find out, then you shouldn't be using Linux anyway. :)
22:17:04 <@duck> which kills kittens
22:17:28 < slart> what about bittornado?
22:17:43 <+Ragnarok> I suppose I could be tempted to write a pygtk gui, but I've got a lot of other stuff to do, and I'm not sure what people want
22:17:45 <+protokol> azureus?
22:17:57 <@duck> part of me is ofcourse making up excuses not to do things
22:18:03 <+protokol> azureus supports plugins
22:18:10 <@duck> protokol: well, write a plugin
22:18:32 <+protokol> heh
22:18:40 < slart> bittornado is based off the offical bt isnt it?
22:18:50 <+protokol> easier said than done
22:18:52 <@duck> slart: I have looked at it and wept
22:19:07 <@duck> it has some improvements, which might be useful
22:19:17 <@duck> but on the other hand it made the whole thing way more complex
22:19:22 <@duck> without cleaning up the original code
22:19:36 <+Ragnarok> gah
22:19:56 <@duck> the GUI feature that you can specify a torrent if no arguments are given is taken from it and added to i2p-bt
22:20:11 < clayboy> let's get the basic bittorrent working excellently before worrying about these fluffy gui things :)
22:20:46 <@duck> slart: probably some other things can be used too; someone just has to do it (properly)
22:21:23 <+ugha2p> clayboy: Well, I think it already does work excellently. :)
22:21:53 < slart> the abc client uses tornado (i think)
22:22:15 < clayboy> i feel like we have still to do some really heavy-duty testing to see how much data can really be pushed through i2p-bt
22:22:21 < bushka> yes it does slart.
22:23:49 <@duck> depending on how those work, you might be able to port the i2p-bt changes to them quite easily
22:24:41 <@duck> please give it a try and report back
22:25:47 <@duck> .
22:25:55 <@duck> any other i2p-bt / bittorrent comments?
22:26:08 < slart> python :S
22:26:41 <+ugha2p> .
22:26:51 <@duck> slart: if you dont like python, you can give porting azureus a try
22:27:00 <+ugha2p> slart: What about it?
22:27:06 < slart> how many people could we get seeding somthing like a linux is for speed testing?
22:27:15 < slart> *iso
22:27:34 <@duck> lets try that after the new i2p release
22:27:57 <@duck> (since pulling an i2p router build from cvs is quite a challenge for most)
22:28:17 <+protokol> eh
22:28:54 <@duck> pl
22:28:57 <@duck> err, ok
22:29:10 <@duck> 3) #idlerpg
22:29:22 <@duck> found this funny irc rpg game
22:29:36 <@duck> you dont have to do anything for it, just idle 
22:29:56 <+ugha2p> Well, you do have to LOGIN. ;)
22:30:04 <@duck> ah ;)
22:30:18 < mule> cvs update -dP :)
22:30:18 < mule> ant dist updater :)
22:30:20 <+postman> it's the most hilarious thing i've ever seen, but i LIKE it :)
22:30:30 <+protokol> there should be prizes
22:30:45 <@duck> on ircnet it has 779 online players 
22:30:46 <+ugha2p> duck: I was thinking, that it could potentially be a reason not to upgrade.
22:30:52 <+protokol> give yodels for winning stuff or reaching levels
22:31:03 <+ugha2p> Although I'm not sure if people on I2P could be that childish. :)
22:31:14 <+protokol> i know duck has like $10000 in yodels
22:31:18 <@duck> protokol: yeah, I have to see how those quests work
22:31:39 <@duck> maybe we can do some fun stuff with it
22:31:42 <@duck> ugha2p: what do you mean?
22:31:49 < ant> * cervantes is not going to do another 40 days without restarting his router
22:32:08 <@duck> ugha2p: oh, not update because of the game :)
22:32:18 <+protokol> Linux: If you can't fix it without restarting, you can't fix it.
22:32:20 <@duck> well, I'll put it on pause while my router restarts
22:32:24 <+ugha2p> :)
22:32:33 <@duck> so if you sync it well, you wont lose
22:32:35 <@duck> hehe
22:32:55 < ant> <cervantes> thats good... since your router restarts all the time :P
22:33:16 <@duck> thats called dedicated testing :)
22:33:20 < ant> <cervantes> I guess that throws roulette into the equation too
22:33:23 <@duck> ok
22:33:38 <@duck> .
22:33:49 <+ugha2p> .
22:34:05 <@duck> 5) ???
22:34:08 <@duck> s/5/4/
22:34:12 <@duck> open mike!
22:34:23 <+postman> .
22:34:53 < mule> with a bit of tweaking you can two routers. one for the game only, which you upgrade only every year
22:34:53 <@duck> questions? comments? suggestions?
22:35:38 < ant> <mahes> Hi, i have a general non-dev question
22:36:08 <@duck> shoot
22:36:08 <+ugha2p> Thanks for holding the meeting, duck.
22:36:50 < ant> <mahes> if i set up an eepsite , how can be reached with an address like i.e mahes.i2p
22:36:59 <+protokol> i have a consern
22:37:44 <+protokol> (start the battle) i think .i2p is a shitty TLD for many reasons
22:38:19 <+ugha2p> mahes: What do you mean 'how'? People will configure their browsers to use the eepproxy, and just enter http://mahes.i2p/ onto their address bar.
22:38:19 <+protokol> i think we should use one that is a) one syllable b) can be pronounced like a word c) does not include a number'
22:38:46 <+ugha2p> protokol: Like .eep?
22:39:07 <@duck> mahes:: to get a 'nice name' to point to your eepsite, it has to be present in your hosts.txt file
22:39:37 <+protokol> ugha2p: sure
22:40:01 <+ugha2p> protokol: You can make a proposal on the mailing list.
22:40:03 <@duck> you can post it on the eepsite announcement forum so others can get it too
22:40:09 <+ugha2p> It'll probably be considered once we have MyI2P.
22:40:35 <+protokol> heh, ill try but jr shot it down for some reason already
22:41:06 < ant> <mahes> well. i am just a user...  ok, so i just publish mahes.i2p=hhfbwer8328... and it will just spread
22:41:32 <@duck> it doesnt spread automatically, ppl need to get it into their hosts.txt somehow
22:41:39 < ant> <mahes> ok
22:41:52 <@duck> but announce it on the forum and it is more likely to :)
22:42:34 <@duck> .
22:43:18 <@duck> lets give it a *baf*
22:43:20 <+ugha2p> .
22:43:30  * ugha2p is waiting for the baffer.
22:43:38  * duck winds up
22:43:45  * duck *baf*s the meeting closed