--- Log opened Tue Aug 03 00:00:27 2021
01:35 < waxwing> undeath (not here): roughly finished reviewing that PR, looks good except i don't get why that recursion is OK, wouldn't it go too deep? (see my comment)
03:57 -!- undeath [~undeath@user/undeath] has joined #joinmarket
04:15 <+JoinMarketRelay> [hackint/nikolai] hello. i've been reading discussion on github related to 'design decentralized network protocol for joinmarket'.
04:17 <+JoinMarketRelay> [hackint/nikolai] i have to agree with belcher where he says - 'This whole plan seems to rely on Tor hidden services right now.'
04:21 <+JoinMarketRelay> [hackint/nikolai] not that i am asked for opinion, but relying on any service that is hard to replace as in change big parts of code and how jm works could bring more issues than good. i still find it easier to take care of and more fair so that every maker and taker can connect to JM network how they want: clearnet, vpn, socks5 (tor,regular proxy..)
04:42 <+JoinMarketRelay> [hackint/nikolai] it's been a while since reading tor v2 spec, but thinking could hidden service descriptors be used to transfer tiny amounts of information/data while smoe parties are offline, this data would then be given out by HS directories. however i do not like the idea relying JM on tor
04:43 < undeath> really all it adds is a mandatory layer of anonymization
04:44 <+JoinMarketRelay> [hackint/nikolai] one big issue i potentially see, and i don't know if v3 fixes this is that 3rd parties can calculate descriptor ids and HS dir fingerprints that will be used in future. imagine for 1 month upfront
04:44 < undeath> make it easy for average users to screw up their privacy and they will
04:45 < belcher_> to be fair joinmarket today relies on tor as well
04:45 < belcher_> using it over clearnet is a really bad idea, vpn is only a tiny bit better
04:45 < undeath> tor is not forced, the irc servers have clearnet acces
04:45 -!- belcher_ is now known as belcher
04:45 < undeath> but yes, nobody _should_ do that
04:45 < belcher> especially for a yield generator
04:46 < undeath> i'm pretty sure users do, though…
04:46 < belcher> "hey this clearnet ip address has loads of bitcoins on a hot wallet!"
04:47 < belcher> to be clear, back a few months ago when tor hidden services went i was pretty disturbed
04:47 <+JoinMarketRelay> [hackint/nikolai] what about pre-generated HS dir fingerprints and HS descriptors used to take down makers or maybe if it would be possible at all to act as them. i see this as potentially big issue. imagine i could become a responsible set of 2x3 HS dirs for a specific onion address. if you go anywhere with tor in future, i think tor onion addrs should be
04:47 <+JoinMarketRelay> changed at least once a week
04:47 < belcher> you have this layer that creates an anonymous internet and now we're reminded it could just go down one day
04:48 < undeath> impersonation is not possible. the signatures in the jm protocol prevent that
04:49 < undeath> and that part about pre-generating some fingerprint or something sounds like FUD right now
04:49 <+JoinMarketRelay> [hackint/nikolai] i would have not known about v3 downtime if it wasn't mentioned in the text
04:49 < belcher> be aware tor gets a lot of FUD around it sometimes
04:50 <+JoinMarketRelay> [hackint/nikolai] did some tests around 2017 related to it, it's not fud
04:50 < undeath> at least for v3 it is
04:50 < undeath> and v2 is deprecated
04:51 <+JoinMarketRelay> [hackint/nikolai] that is true, untested in v3
04:55 <+JoinMarketRelay> [hackint/nikolai] all in all, with that saying i wanted to say if you ever choose tor to carry bones of JM, do not let users constantly use same onion addrs, but maybe regenerate them every now and then, like once a week at most. back in 2017 testing somebodies else idea we've came to conclusion that it is possible to take down onion services for the time while
04:55 <+JoinMarketRelay> specific fingerprints of HS dirs are responsible for specific onion address. this attack would have to be plan
04:56 <+JoinMarketRelay> [hackint/nikolai] s/lays/for relays/
04:58 < belcher> that shouldnt be too hard to do
05:05 <+JoinMarketRelay> [hackint/nikolai] but in case this still works for v3, it also means having a well known longterm hardcoded onion addr in a config that is for a directory server as called in the github issues would also be prone to such attacks, maybe even more if it was dedicated server. for the moment using other people irc servers might be saving JM from hddos.
05:06 <+JoinMarketRelay> [hackint/nikolai] i will think about this, i find it interesting. have a good day.
05:32 < belcher> im looking at the sybil resistance page of ob-watcher, right now if a fidelity-bond-aware taker does a coinjoin with 6 makers a sybil attacker would have to lock up 19303 BTC for 1 year or burn 24 BTC to successfully attack
05:46 <+JoinMarketRelay> [hackint/nikolai] beautiful!
05:47 <+JoinMarketRelay> [hackint/nikolai] i wanted to ask earlier, maybe i've missed it. what's the intended way of sending to FB address. extract priv keys and sweep using bitcoin-cli to FB address?
05:49 <+JoinMarketRelay> [hackint/nikolai] since i haven't seen in sendpayment.py -h that it can sweep specific address, only specific mixdepths
05:49 < belcher> id use sendpayment and coin control (i.e. freezing utxos you dont want to send)
05:51 <+JoinMarketRelay> [hackint/nikolai] that sounds easy. appreciate!
06:05 < belcher> the "6 makers" actually doesnt take into account the bondless maker allowance
06:06 < belcher> which by default is 12.5% or 1-in-8, so that 6 makers would on average be 6/(1-0.125) = 6.857
06:07 < belcher> so the more correct statement (with rounding) is if a fidelity-bond-aware taker does a coinjoin with 7 makers then same thing
06:08 < belcher> kristapsk and others: https://twitter.com/joinmarket/status/1422543725843070978
06:08 < belcher> https://x0f.org/@belcher/106692349206727454
06:08 -!- jungly [~jungly@gateway/tor-sasl/jungly] has joined #joinmarket
06:09 -!- jungly [~jungly@gateway/tor-sasl/jungly] has quit [Remote host closed the connection]
06:14 -!- common [~common@user/common] has quit [Quit: Leaving]
07:09 < waxwing> kristapsk .. wait, are you saying in Latvian you use apostrophes for plurals? it seemed to me that in Latvian you just put an 's' at the end of basically everything :)
07:40 -!- davterra [~davterra@143.198.56.186] has joined #joinmarket
07:42 -!- tralfaz [~davterra@143.198.56.186] has joined #joinmarket
07:43 -!- davterra [~davterra@143.198.56.186] has quit [Killed (NickServ (GHOST command used by tralfaz))]
07:43 -!- tralfaz is now known as davterra
08:35 <+JoinMarketRelay> [hackint/kristapsk] waxwing, we do use apostrophs when you are trying to conjugate non-Latvian words, UTXO could be one of them
08:35 <+JoinMarketRelay> [hackint/kristapsk] but, yes, putting s and the end of everything is our thing :)
08:37 < waxwing> yeah it's a weird thing. As an example, you would *never* see someone write 1970's about 20 years ago. When people did, it was an error. Nowadays there are plenty of people who don't think that's wrong.
08:38 < waxwing> That's language for you :) What is right is dictated by errr consensus, eventually :)
08:38 <+JoinMarketRelay> [hackint/kristapsk] just remembered one example from 1942 German occupation magazine, proper Latvian would be "Ādolfa Hitlera", but there they used German names, so it was "Adolf Hitler'a"
08:48 <+JoinMarketRelay> [hackint/nikolai] heh. Adolf -> dolFa way of translating words sounds like satrovacki in some balkan ex-yu countries. :)
08:49 <+JoinMarketRelay> [hackint/nikolai] like govno, vnogo, or mojne me batje = nemoj me jebat. do you do that also?
09:03 < waxwing> the README-linked http://qvzlxbjvyrhvsuyzz5t63xx7x336dowdvt7wfj53sisuun4i4rdtbzid.onion/obwatcher/ is a good example of why i kept being "don't bother" about public orderbookwatchers: it shows 42 in the orderbook.
09:03 < waxwing> fixing #769 is better than publishing info that's wrong (albeit not particularly harmful!)
09:50 <+JoinMarketRelay> [hackint/user_] @waxwing aren't you looking at the old non-native sw orderbook? I see 40ish orders there and over 200 hundred in the native sw one
09:59 -!- SomeFellow [~SomeFello@194.54.28.203] has joined #joinmarket
11:20 < waxwing> user_ doh! of course you're right. well that's a lot better :)
11:20 < waxwing> i'm actually not sure how many offers get lost over what time rate, but refreshing it regularly "kinda" solves the problem. as long as no one treats it as gospel.
11:32 -!- undeath [~undeath@user/undeath] has quit [Quit: WeeChat 3.1]
15:39 < xyy> was building JM on top of I2P ever considered? 
15:39 < xyy> curious to see the discussions 
15:41 < xyy> and the very light version of that, were IRC networks that are reachable over I2P considered? 
15:41 < xyy> just looking at the default `tunnels.conf`, they have `irc.postman.i2p` and `irc.ilita.i2p`
15:42 < xyy> (so I'm sure there are more ...)
16:47 -!- SomeFellow [~SomeFello@194.54.28.203] has quit [Quit: Connection closed]
17:11 -!- JMBridged [~CoinJoins@gateway/tor-sasl/jmbridge] has joined #joinmarket
17:12 -!- JMBridge [~CoinJoins@gateway/tor-sasl/jmbridge] has quit [Ping timeout: 244 seconds]
17:12 -!- JMBridged [~CoinJoins@gateway/tor-sasl/jmbridge] has quit [Remote host closed the connection]
17:13 -!- JMBridge [~CoinJoins@gateway/tor-sasl/jmbridge] has joined #joinmarket
17:14 -!- belcher_ [~belcher@user/belcher] has joined #joinmarket
17:17 -!- belcher [~belcher@user/belcher] has quit [Ping timeout: 240 seconds]
22:01 -!- JMBridged [~CoinJoins@gateway/tor-sasl/jmbridge] has joined #joinmarket
22:01 -!- JMBridge [~CoinJoins@gateway/tor-sasl/jmbridge] has quit [Ping timeout: 244 seconds]
22:01 -!- JMBridged is now known as JMBridge
23:34 -!- belcher_ is now known as belcher
--- Log closed Wed Aug 04 00:00:28 2021