IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#ls2
/2022/01/10
orignal inetersting
orignal I'm kicked out too frequentl
zzz 0) Hi
zzz happy new year
orignal you too
zzz whats on the list for today?
orignal nothing from me yet
zzz I'll add 1) SSU2
zzz ok guess that's it for now
zzz 1) SSU2
zzz I haven't thought about it in a month at least
zzz so what I want to do is make a list of what is still TODO
zzz so we can finish up the spec and get to coding pretty soon
zzz So I'd like to work on that list and bring it next week
zzz then we can split up the work and get it finished
zzz sound good?
orignal we should start from handahsking I guess
zzz I believe the handshake part is done, but the KDFs probably are not
zzz but I can't remember
zzz as usual, takes a while to get it all in my head
zzz anything else on 1) ?
orignal if it's done can we strart coding?
zzz sure
zzz there's two ways to find out what's NOT done:
zzz 1) review it
zzz 2) start coding ))
orignal I prefer 2
zzz your choice
zzz anything else on 1) ?
zzz I'll be back with my list of what's not done next week
eyedeekay1 Well it isn't done is it?
eyedeekay1 I've been giving it a lot of homework time and it seems like, for instance, KDF's largely say "exactly as noise" which might be fine, but since I'm catching up to everybody else on what properties that provides, I actually have no idea whether it's done
zzz it's certainly not done
zzz but there's really 3 parts:
zzz 1) basic connections
zzz 2) peer test
zzz 3) relay
zzz 1) should be pretty close except for the KDFs
zzz so maybe if we wanted we could start coding on 1) before finishing the spec for 2( and 3)
orignal what with KDF?
orignal shouldn't it be the same as for NTCP2?
zzz still have to do a split() and more HKDF
orignal what for?
zzz for the data phase keys
zzz I think
zzz but again, I'll be smarter next week
orignal but we do split for NTCP2
orignal for _ab and _ba
zzz yeah but the data phase is very different... no siphash...
orignal because we don't need to pass length
zzz right
orignal it's in UDP header
orignal and I guess we can use padding up to MTU
zzz padding blocks, same as NTCP2, yes
zzz eyedeekay1, you think you might try to code along with us in Go?
orignal not the same
orignal because we are free to fill it up to MTU size
orignal with no cost
orignal I'm even not sure if we need it at all
zzz yes, but you want the padding covered by AEAD
zzz so we don't want padding outside the blocks
eyedeekay1 I can try, would be easiest to follow if you could put it on a branch for me
zzz there's nothing to put, we would all be working in parallel
eyedeekay1 OK will try to keep up :)
zzz would be a lot of work and you probably have a lot more todo just to have any transports at all
zzz anyway
zzz I thought of a 2), so anything else on 1) ?
eyedeekay1 Yeah but I can at least try, maybe learn where things go and write comments where I don't have something yet
orignal what's the proposal # ?
zzz and speaking of proposals
zzz 2) UDP Trackers
zzz I wrote this over the holidays
zzz I emailed ot to R4SAS since he's working on opentracker
zzz I don't think orignal cares, but maybe?
zzz we can discuss it here, or not
zzz just FYI for now
zzz I've already gotten feedback from BiglyBT
orignal opentracker is a replacement of postman
orignal if he goes down again
orignal will check 160
zzz we also have ZzzOT which is a plugin only
zzz I could do ZzzOT standalone if somebody wanted it
orignal what's that?
zzz opentracker can't ever be a replacement of postman, it's just an opentracker
zzz ZzzOT is a java open tracker in a console plugin
zzz based mostly on the i2psnark code
zzz very simple
zzz I wrote it in March 2010
zzz a few people run it
zzz anything else on 2) ?
eyedeekay1 I've read it, so far my only commentary is that some of it seems "Optional" so to speak? Like the "fast mode"
zzz it's all optional, because we don't support it now
zzz there's two modes, one that's easier if your code supports clearnet udp trackers now, and one that's easier if you're coding it from scratch
zzz the former will be BiglyBT; the latter will be i2psnark
eyedeekay1 Yeah that's kind of what I was getting at
eyedeekay1 What I'm wondering is if there's an effective way to detect whether one or the other is supported and choose it?
eyedeekay1 And if so, wouldn't that give away your bittorrent client?
zzz when you announce, you 'give away your bittorrent client' anyway
eyedeekay1 Yeah I guess that's not something a client would spoof so the point is moot. Nevermind, thinking about it like web fingerprinting
zzz anyway, just FYI for now
zzz anything else on 2) ?
zzz anything else for the meeting?
orignal busy with kazahstan affairs ))
zzz ok, good luck with KZ
zzz thanks guys
eyedeekay1 Thanks zzz
orignal i2p over ygg works good for them
zzz nice
orignal they connect ygg to that ports
orignal some people run outside KZ including me
orignal and then run i2p over this link
eyedeekay1 What do you do to handle/configure your yggdrasil links?
eyedeekay1 Are you simply making I2P connections over both the clearnet and ygg links or is it more elaborate than that?
orignal yggdrasil is another ipv6 adress
orignal from 200-399 range
orignal so you you support ygg and your peer has ygg address in RI
orignal link can be established through ygg
orignal more interesting if your router doesn't have any transports but ygg
orignal in this case you must make sure that first hop supports ygg
orignal same for floodfill
orignal usually you publish yourself through eploratory tunnels
orignal because you can't reach a FF directy
orignal another issue is reseed
orignal ygg-based reseeds always include few routers with ygg
RN does each hop in a tunnel need to support ygg?
orignal only next to you
RN fun stuff
orignal why? it just works