I2P dev meeting, January 15, 2013 @ 20:00 UTC

Quick recap

  • Present:

    darrob, dg, eche|on, hottuna, KillYourTV, LaughingBuddha, orion, sponge, welterde, zzz

  • Next Meeting

    The next meeting is scheduled for Tuesday, January 22 @ 20:00 UTC (8:00PM)

完全な IRC ログ

20:02:07  <hottuna> dg, is there a meeting today
20:02:26  <LaughingBuddha> oh yeah, it's tuesday
20:02:56  <hottuna> str4d, Meeh: is the meeting now?
20:03:49  <orion> yes
20:04:07  <Meeh> hottuna: dunno, hope so
20:04:26  <hottuna> alright, is there an item list?
20:04:56  <hottuna> dg, ping
20:06:07  <hottuna> alright, lets start it in 5 minutes
20:06:35  <hottuna> everyone around, igure out what you would like to talk about
20:06:57  <hottuna> when we start help create the item list
20:07:36  <iRelay> <weltende@freenode> I though today was the crypto meeting
20:08:07  <hottuna> weltende, i missed the last one. Im just trying to create some order
20:08:22  <hottuna> zzz, crypto meeting
20:10:23  <hottuna> weltende, what crypto was going to be discussed?
20:11:09  <hottuna> Alright, lets try to start this thing.
20:11:16  <hottuna> Meeting schedule:
20:11:28  <hottuna> * Syndie hackathon status
20:11:51  <hottuna> * I2P crypto
20:12:02  <iRelay> <weltende@freenode> hottuna: that I do not know..
20:12:20  <iRelay> <weltende@freenode> hottuna: should I send your r0ket per package or something btw?
20:12:32  <dg> am I in time?
20:12:51  <hottuna> i just started, but if you'd like to take over i'd be happy
20:12:53  <iRelay> <weltende@freenode> dg: yup
20:13:20  <dg> the most I know is "crypto"
20:14:32  <hottuna> dg, any other items we should take care of
20:14:35  <hottuna> ?
20:15:14  <dg> Syndie status, any updates from the researchers, general direction?
20:15:37  <hottuna> alright. lets start then?
20:15:55  <dg> sure
20:16:14  <dg> I'm hoping zab can appear as him and zzz know the most about the router internals..
20:18:28  <orion> I think we were going to discuss possibly upgrading the crypto.
20:18:28  <hottuna> moving away from dsa-1024 anyway
20:18:31  <orion> And the logistics behind that.
20:18:31  <hottuna> alright
20:18:31  <dg> Let me dig up the topic.
20:18:31  <hottuna> that makes sense
20:18:31  <hottuna> dg, http://zzz.i2p/topics/715
20:18:31  <dg> a-ha
20:18:42  <dg> ha.. two years ago
20:19:00  <hottuna> Should we start with "Syndie status" for the time being?
20:20:11  <dg> Yeah
20:21:09  <hottuna> Alright, how many items on the todo-list have been added / completed?
20:21:40  <hottuna> And whats the url for the todo?
20:25:43  <hottuna> Syndie changelog: http://zzz.i2p/posts/6095
20:25:43  <hottuna> Are there any syndie users around?
20:26:06  <dg> I've tested it but I've had issues/found it dead.
20:26:09  <hottuna> I tried it and started synching, but there are a lot of items to synch
20:26:16  <dg> Not sure if anyone is a regular user..
20:26:37  <hottuna> and if you allow a long histoty, maybe the newest posts should be fetched first?
20:26:44  <darrob> i'm kinda here.
20:27:08  <hottuna> darrob, you run an archive as well?
20:27:16  <darrob> hottuna: yes
20:28:23  <hottuna> have you been following the hackfest?
20:29:25  <hottuna> Syndie todo: https://github.com/i2p/i2p.syndie/blob/master/TODO
20:29:36  <eche|on> syndie hackfest was a success and it works quite well. currently there is a small issue with to much posts for first starters.  Will resolve soon.
20:30:28  <hottuna> great is there a trac item for large initial synch or is it noted somewhere else?
20:30:32  <eche|on> syndie fetches by date
20:30:43  <darrob> yes, the hackfest is a huge success so far. zzz has fixed a whole bunch of bugs that made using syndie hard. even the pulling in of old messages should be better now, iirc.
20:30:46  <eche|on> and there are some discussions about "what is new"
20:31:34  <eche|on> in some kind of anonymity reason "all the archive gets today is NEW"
20:31:59  <hottuna> eche|on, I dont understand that last part. Are fetches done by latest post first?
20:32:05  <eche|on> no
20:32:13  <hottuna> When I synched recently it seemed pretty random
20:32:17  <eche|on> imho it fetches one after another, random
20:32:31  <eche|on> but it sends all messages which are "new" since a special date
20:32:42  <hottuna> maybe some order would be preferable
20:32:49  <eche|on> but new is/was not based on message creation date
20:33:04  <eche|on> new was based on date of arrival on server (out of anonymity reasons)
20:33:14  <hottuna> aah, I see
20:33:48  <hottuna> where should I go to suggest ordered fetching?
20:34:13  <eche|on> and if a server gets 5000 news from some "lost old archive" on one day, those 5000 are seen as new, even if the creation date is >5 years ago
20:34:16  <darrob> hottuna: post it in "Syndie dev". :)
20:34:27  <eche|on> or snydie bug  ;-)
20:34:34  <hottuna> the syndie thread?
20:34:49  <zzz> we spent a week working on pushes, almost done, have to work on some date issues, then will look at pulls
20:35:08  <darrob> no, the syndie forum.
20:35:11  <eche|on> hottuna: oh, not yet in syndie -)
20:35:26  <eche|on> but you can change the default pull/push policy in your syndie
20:35:45  <hottuna> I've seen some settings but Im a pretty novice user
20:36:14  <zzz> but main point: I'm moving back to i2p, need you all to step up with more testing and coding
20:36:14  <eche|on> pull policy
20:36:29  <hottuna> ok, that is good to know zzz
20:36:44  <eche|on> yeah, I was under the impression zzz pulled the "jrandom card" and moved from I2P to syndie ;-)
20:36:44  <hottuna> Any other comments re syndie?
20:37:10  <hottuna> Alright, moving on
20:37:18  <hottuna> ** Updates from the researchers **
20:37:34  <hottuna> zzz, ahve you had any contact with any of them?
20:37:40  <zzz>  will try to pick a spot for a release, probably numbered 1.103b-0, will need help from weltende to get it on website
20:37:55  <sponge> Question for everyone: Who here has a google account, and has balls of steel, and can place android related i2p stuff on a google drive account to allow faster access to android i2p related stuff? --> *IDEA* if i2p already HAS a google account, who is in charge of it, so I can get stuff to them to publish it on google drive?
20:38:06  <eche|on> hottuna: btw, we missed the cool mathematics beside us explaining elliptic curves
20:38:23  <eche|on> sponge: we do habe the I2P google account?
20:38:30  <eche|on> sponge: see query
20:38:30  <zzz> only what you saw here and on the zzz.i2p thread. but if christoph has anything to add, he is here
20:38:51  <hottuna> eche|on, I didnt miss it. I've been through that before. I had flashbacks and the sweats..
20:38:51  <sponge> eche: that's the question... do we?
20:39:02  <zzz> sponge, maybe stick to meeting topic?
20:39:05  <hottuna> christoph, we're having the weekly i2p meeting now
20:39:16  <hottuna>  do you have any i2p-related updates?
20:39:19  <sponge> oh meeting? excellent!
20:39:59  <orion> Google account?
20:40:02  <orion> I'll do it.
20:40:14  <hottuna> Related to the research they're doing I read the GNUnet dht paper
20:40:21  <sponge> orion: Wow! xcellent!
20:40:24  <hottuna> and it was interesting
20:40:27  <eche|on> hottuna: lovely sweets, I was just impressed about those two crazy guys and I think elliptic curves are VERY cool now^^
20:40:54  <eche|on> sponge: what google accoutn do you need?
20:41:00  <sponge> zzz/others: I'm not a crypto expert, I have nothing to contribute to the meeting :-(
20:41:27  <dg> me too
20:41:40  <hottuna> R5N,  gives nodes doing a lookup the option to just continue looking and having it eventually work as long as there is a single path to the desired key
20:42:00  <eche|on> sponge: and please see the query
20:42:19  <hottuna> It think it might be an interesting alternative for the netdb
20:42:30  <sponge> eche: basically someone who will be able to put up a large binary and apk's for android, including i2p-android (zzz's and the dev build) plus the libc stuff I am working on, which will be an apk file and a tarball
20:42:45  <hottuna> but im sure zzz isn't very keen on redoing the dht just quite yet :P
20:43:31  <eche|on> sponge: we got a I2P google code account. which could host that files, to. I just cannot tell about google play account
20:43:54  <hottuna> Alright, no further updates on the research?
20:44:13  <sponge> eche: the idea is to have something easy for people to get the APKs and tarball easily and as trouible-free as possible, instead of over i2p. It has nothing to do with Play store, Google Drive is free unless you want more space IIRC...
20:44:20  <zzz> sponge, maybe stick to meeting topic?
20:44:23  <hottuna> Moving on...
20:44:42  <sponge> will talk about it later
20:44:45  <hottuna> **  I2P crypto **
20:44:52  <eche|on> sponge: ah, ok, we do have official I2P google code hosting platform. Thats the best place IMHO
20:44:52  <zzz> a+ for starting mtg almost on time; step 2 is keeping it on topic and moving :)
20:45:19  <sponge> eheheh, I just happened to be awake ;-)
20:45:26  <hottuna> So what specific crypto decisions do we have infront of us?
20:45:41  <hottuna> Replacing DSA 1024
20:45:48  <zzz> I'd like to know who put "crypto" as the mtg topic, and what in particular did you have in mind?
20:46:09  <hottuna> I was told by welt
20:46:28  <dg> str4d but I'm not sure he's here.
20:46:39  <sponge> I say replace DSA 1024 with what ever needs to be done, and we need a way to still allow older routers to recognize what is in-use now.
20:46:50  <hottuna> No harm discussing it
20:47:20  <zzz> its a big topic of course, I'd like to hear what and why it's on the agenda, so we can stay focused.
20:48:06  <hottuna> this is why we should have a forum thread for every meeting
20:48:24  <eche|on> syndie *g*
20:48:32  <sponge> I think the largest mistake made in I2P was in not looking far forward enough to have some kind of hint as to what kind of crypto to use. Perhaps what could be done is to offer a hint via DSA 1024 in such a way that the older router will ignore such a hint, and a newer one will take the offer
20:48:55  <hottuna> thats is a good point sponge
20:48:58  <zzz> well (OT) we need somebody in charge of every meeting, so it has an agenda and can start on time. dg, are you going to start doing that again? if not we need to find somebody else
20:49:01  <orion> 20:21:47<+str4d> KillYourTV: could you update the /topic for the crypto meeting next week?
20:49:01  <orion> 20:22:57-!- KillYourTV changed the topic of #i2p-dev to: Syndie hackfest in progress, current Syndie build 1.102b-3 | Current I2P build: 0.9.4-4 | Dev Crypto mtg here 8:00 PM (20:00) UTC Tues. Jan. 15 | Report bugs at http://trac.i2p2.i2p / http://trac.i2p2.de
20:49:07  <iRelay> Title: I2P (at trac.i2p2.i2p)
20:49:27  <dg> Yeah, I'm fine with that. I got out of sync after I died. :)
20:50:15  <hottuna> Alright, I dont think the crypto bit is leading anywhere. Objections?
20:50:18  <sponge> zzz: is it possible to offer a hint in DSA 1024 in such a way that current and older routers will ignore?
20:51:02  <dg> hottuna: sadly, we pinned that as the topic and I didn't spend any time thinking of anything else..
20:51:08  <zzz> dg, ok with which? doing it again or having somebody else do it?
20:51:19  <sponge> will a 'dual crypto stack' cause much bloat?
20:51:22  <dg> We can talk about where we're going with things and our plans for restricted routes if you guys fancy it. It seems important to me but..
20:51:26  <dg> zzz: doing it again
20:51:53  <KillYourTV> meeting threads in Syndie <3
20:52:36  <hottuna> sponge, i think there are complications related to having multiple cryptos
20:52:43  <zzz> sponge, very briefly, for each _use_ of crypto (e.g. sigining RIs), not for each _type_ of crypto (e.g. DSA 1024), there's two things to do: decide if and when we need to change, and decide how to do it in a compatible way if possible. I've started the latter, and the thread's on zzz.i2p.
20:52:58  <hottuna> like crypto-downgrading attacks etc.
20:53:09  <zzz> but we can't talk about "DSA 1024", only a particular use of it (there's at least half a dozen iirc)
20:53:43  <zzz> and if we are going to talk about upgrading crypto next week, the tor paper is required reading in advance.
20:54:18  <zzz> (search for "read it and weep" on zzz.i2p)
20:54:28  <hottuna> dg, could you start a meeting thread on zzz.i2p or forum.i2p?
20:54:47  <hottuna> so that some notes for next week can be arranged?
20:55:35  <hottuna> link to tor paper:  https://gitweb.torproject.org/tor.git?a=blob_plain;hb=HEAD;f=doc/spec/proposals/ideas/xxx-crypto-migration.txt
20:55:38  <dg> hottuna: http://zzz.i2p/topics/1268 - unless we need a new one?
20:56:12  <sponge> one thing we don't need is a I2P flag day
20:56:36  <dg> for the love of god no
20:56:39  <zzz> so sponge, in general, "bloat" is not the issue. compatibility is. pls read the read-it-and-weep paper
20:56:42  <hottuna> doh, that link is dead
20:56:42  <hottuna> actual tor link: https://gitweb.torproject.org/torspec.git/blob_plain/34ecac0fbac7f476bfcbf813767721fada62c17e:/proposals/ideas/xxx-crypto-migration.txt
20:56:48  <LaughingBuddha> hehe
20:56:59  <sponge> yes, i did read that thread a while ago
20:57:03  <hottuna> alright, any more comments on this topic?
20:57:41  <hottuna> dg, maybe we should have one per week so that topics can be discussed
20:57:58  <hottuna> and a topic list can be created
20:58:01  <sponge> as long as it is a smooth transition, just do what is needed, I'll be happy with it
20:58:08  <dg> might be a good idea for keeping things fresh but could get annoying over time
20:58:36  <hottuna> dg, we could at least try?
20:58:43  <hottuna> and see if it works out
20:58:54  <sponge> also, am I correct in that it won't change the b32.i2p format?
21:00:28  <sponge> i.e. it's still a hash
21:00:35  <dg> sure
21:01:47  <orion> sha256 seems reasonable for now
21:02:02  <hottuna> destinations relate to public/private keys
21:02:02  <hottuna> as long as we dont switch pub-keys or hashes we should be fine
21:02:37  <orion> What if we used an ECC primitive that resulted in public-key lengths which are the same that we use now?
21:02:44  <orion> i.e, 128 bytes
21:03:06  <hottuna> orion, that would be a terrible wast of cpu-cycles and bandwidth
21:03:10  <eche|on> I think those are 156 bits
21:03:29  <hottuna> and we still would need to notify the receiving end of what crypto we are using
21:04:42  <hottuna> alright, any more discussion relating to crypto?
21:05:34  <hottuna> Moving on...
21:05:44  <sponge> :-)
21:05:47  <hottuna> ** Open floor / Misc **
21:06:22  <eche|on> i cleaned ugha.i2p
21:06:37  <hottuna> there was a lot of spam around?
21:07:03  <eche|on> oh yeah, a few hundred spam articles
21:07:26  <hottuna> It's really nice that ugha.i2p is still around
21:07:44  <hottuna> I've read and posted some helpful stuff there
21:08:15  <hottuna> dg, could you create the meeting thread for next week?
21:08:31  <hottuna> orion, how is i2pcpp coming along?
21:08:42  <dg> Yeah. Any thing you want me to mention or just a "This is a meeting thread for discussion of $date's meeting.."?
21:08:44  <orion> hottuna: Inbound SSU works now.
21:09:07  <orion> I also can begin the build of inbound tunnels.
21:09:07  <hottuna> wow, you're pretty fast :)
21:09:19  <orion> Yeah, it's coming along very quickly.
21:09:34  <hottuna> dg, nothing special I was thinking of jumpstarting the crypto talk
21:09:34  <orion> College starts up again on Monday though.
21:09:37  <sponge> yea orion is pretty fast, considering how horribly obtuse c++ can be
21:09:47  <orion> Things are going to slow down a LOT after Monday.
21:10:02  <orion> So I am trying to get as much done as possible.
21:10:02  <dg> hottuna: Alright, will sum up the dilemma.
21:10:13  <LaughingBuddha> Maybe someone could create a page @ ugha with stuff that should be added
21:10:28  <hottuna> naturally, but there is a atleast a codebase we can direct people complaining about java to :P
21:10:51  <hottuna> LaughingBuddha, what stuff added to what? i2pcpp?
21:11:06  <LaughingBuddha> Sorry, stuff that we should have on the wiki
21:11:13  <LaughingBuddha> Like a request page
21:11:16  <hottuna> aah :)
21:11:19  <hottuna> good idea
21:11:19  <LaughingBuddha> Tutorials, etc
21:11:42  <LaughingBuddha> Would be nice to have a site to point new people to for common questions
21:11:53  <LaughingBuddha> How do I set up X
21:11:56  <orion> When is the new site rolling out?
21:12:47  <hottuna> orion, last time i was talking to str4d he was unhappy with a few of the pages
21:13:06  <orion> I see.
21:13:25  <hottuna> but Im thinking that maybe just launching it and working out small issues might be a good idea
21:14:08  <hottuna> eche|on and weltende: how much work is it for you to migrate to the new site?
21:14:34  <iRelay> <weltende@freenode> hottuna: not too much.. is it ready already?
21:14:57  <hottuna> I think it's getting there, str4d is no quite happy yet
21:15:22  <hottuna> but with str4d's blessing I would like to launch it before it is perfect
21:15:49  <sponge> living documents are never 'perfect'
21:16:16  <hottuna> that is my thinking
21:16:16  <sponge> so, you mean 'good enough'
21:16:18  <LaughingBuddha> :)
21:17:03  <hottuna> I think it is good enough. And a bit better than the current one
21:17:55  <LaughingBuddha> a bit?
21:18:06  <hottuna> quite the bit
21:18:21  <LaughingBuddha> It's like putting the old site in a time machine
21:19:48  <dg> anything apart from crypto, hottuna?
21:20:03  <hottuna> not that i can think of
21:20:19  <dg> alright, so we're going to discuss the DSA .. again?
21:20:48  <hottuna> like zzz, whe need to discuss every place where every cipher is used
21:21:11  <sponge> my thoughts and concerns as to DSA/crypto have been expressed, I have nothing more to say
21:21:49  <dg> hopefully can get zab on by then?
21:22:08  <hottuna> that would be optimal
21:22:23  <dg> hottuna: this ok? http://zzz.i2p/topics/1328
21:22:26  <hottuna> LaughingBuddha, I've added a requests section to http://ugha.i2p/
21:22:29  <iRelay> Title: zzz.i2p: Meeting [22nd January] (at zzz.i2p)
21:22:32  <dg> I'll email zab now.
21:22:39  <hottuna> Great dg!
21:22:42  <hottuna> thank you
21:22:45  <sponge> one thing to note about meetings when it concerns coders--- ever tried to hurd cats?
21:22:52  <hottuna> alright, anything else before I close this meeting?
21:22:59  <LaughingBuddha> Okay hottuna
21:23:10  <dg> hottuna: I think I should include getting ugha updated and talking about what we need to finish before we put the new site live
21:23:10  <hottuna> sponge, that's the feeling Im getting
21:23:13  <dg> I don't think so
21:23:31  <hottuna> agreed
21:23:46  <hottuna> Alright. This meeting is done.
21:23:52  <hottuna> See you next week, same time.
21:23:52  <hottuna> baf's
21:24:12  <dg> huzzah