--- Day changed Wed Nov 16 2016
00:02 -!- rdymac [uid31665@gateway/web/irccloud.com/session] has joined #joinmarket     
00:02 -!- rdymac [uid31665@gateway/web/irccloud.com/session] has quit [Changing host]     
00:02 -!- rdymac [uid31665@gateway/web/irccloud.com/x-wdzsyrbkpxnremas] has joined #joinmarket     
00:02 -!- rdymac [uid31665@gateway/web/irccloud.com/x-wdzsyrbkpxnremas] has quit [Changing host]     
00:02 -!- rdymac [uid31665@unaffiliated/rdymac] has joined #joinmarket     
00:02 -!- rdymac [uid31665@unaffiliated/rdymac] has quit [Changing host]     
00:02 -!- rdymac [uid31665@gateway/web/irccloud.com/x-wdzsyrbkpxnremas] has joined #joinmarket     
00:03 -!- lnostdal [~lnostdal@62.90-149-73.nextgentel.com] has quit [Ping timeout: 268 seconds]     
00:17 -!- lnostdal [~lnostdal@62.90-149-73.nextgentel.com] has joined #joinmarket     
01:10 -!- Yohkii [~Yohkii___@unaffiliated/yohkii] has joined #joinmarket     
01:15 -!- Yohkii [~Yohkii___@unaffiliated/yohkii] has quit [Quit: Leaving]     
01:23 -!- DeathShadow--666 [~IDSE@S0106a84e3f595813.vc.shawcable.net] has quit [Ping timeout: 256 seconds]     
01:25 -!- DeathShadow--666 [~IDSE@S0106a84e3f595813.vc.shawcable.net] has joined #joinmarket     
02:14 -!- rdymac [uid31665@gateway/web/irccloud.com/x-wdzsyrbkpxnremas] has quit [Quit: Connection closed for inactivity]     
02:22 -!- rdymac [uid31665@gateway/web/irccloud.com/x-fxdugzizveayzhtq] has joined #joinmarket     
03:15 <@waxwing> https://www.reddit.com/r/Bitcoin/comments/5d4urg/circle_is_proactively_monitoring_the_origin/  <- unfortunately there are no details as usual, unfortunately by the nature of the thing there probably can't be (usually reasons are not given for account closure)
03:21 < JM-IRCRelay> [AlexCato1] noticed: maker "J59YbBMtQeErbBeo" seems to copy every existing offer on the orderbook (same cj fee, tx fee and max offer size for absoffers, 124btc for reloffers)
03:23 <@waxwing> huh, that's a new one. except for the annoyance of extreme spamminess, that doesn't really break rules or gain unfair advantage. unless i missed something.
03:24 <@waxwing> what's the betting it's mr 7.49 trying a new method? :)
03:28 <@waxwing> well, it overlaps amount ranges, which we avoided to not be spammy, the taker code defends against that, so if whoever coded that could re-code it not to overlap that would be nice, but not something to lose sleep over, right.
03:32 <@waxwing> oh my mistake, 7.49 is still there, also i think there are some other duplications
03:34 <@waxwing> anyone has any ideas about this, please do pursue them :) best i can think of is something like a p2ch deposit, i can't currently imagine a way of proving distinct utxo sets that makes any sense or is practical.
03:34 <@waxwing> in other news, crossed 10k hashes in my blacklist yesterday.
03:35 < JM-IRCRelay> [AlexCato1] yeah, just ran a test sendpayment to study the debug log. Maker code seems to be fine, and the 7.49 dude indeed is still there,  and probably still makes the first one or two join-tries unsuccessful with his sybil nicks (if they still dont answer, i did not check that)
03:36 < JM-IRCRelay> [AlexCato1] *taker code (not maker)
03:36 <@waxwing> "still makes the first one or two tries unsuccessful"; even with minimum_makers=2?
03:37 <@waxwing> cos i don't find that nowadays. i usually do something like -N 6 (i.e. similar to default) and it goes through, but maybe the situation is getting worse?
03:37 < JM-IRCRelay> [AlexCato1] possibly, but didnt do the math. They are an order of magnitude cheaper than other offers (e.g. for a 5btc join, they ask for 1000sat, while all others ask for 10,000+sat)
03:38 <@waxwing> yes, that's true, i usually see them in there.
03:39 <@waxwing> since way back i've been thinking default order chooser should be "random under a max" not weighted, because of this potential attack vector.
03:39 < JM-IRCRelay> [AlexCato1] i'd have to look in detail at the weighting, just how much effect the cj fee has on the weighting
03:39 <@waxwing> which is now not so much potential as actual :)
03:39 <@waxwing> it's just an exponential distribution.
03:39 <@waxwing> of course one can/should just -P but that is too much cognitive load on users.
03:39 <@waxwing> and doesn't work for tumbler anyway.
03:40 < JM-IRCRelay> [AlexCato1] so an offer with cjfee=1sat is infinitely more likely to be chosen than an offer with cjfee=1000sat?
03:40 <@waxwing> random under a max weakens the price signal, which is unfortunate, but it counters the "swamp the bottom of the orderbook" attack
03:40 <@waxwing> no exponential doesn't result in infinitely high probability at zero.
03:41 <@waxwing> any prob. distn has to be normalized (total P=1)
03:42 <@waxwing> well non-sequitur there (unbounded region can still be finite :) but in any case, yeah it doesn't blow up in that way.
03:42 < JM-IRCRelay> [AlexCato1] yeah, of course. But with 5 offers at 1sat and others at 1k+sat, does the distribution look more like 19,9% for the five 1sat-offers and (practically) nothing for the others
03:42 < JM-IRCRelay> [AlexCato1] ah, last sentence answered that already, thanks
03:42 < JM-IRCRelay> [AlexCato1] indeed need to plot the weighting function once to get an intuitive feeling for it
03:42 <@waxwing> yeah i remember wondering this myself and looked at it, it was fine, it gives sane results. i vaguely remember 1/e being involved somehow :)
03:43 <@waxwing> 1/e is always involved somehow. i remember once being asked to guess the value of some weird expression and i guessed 1/e and was right. so basically 1/e is usually a good guess for things :)
03:45 <@waxwing> along the same lines, we should delete -C from the codebase imo.
03:46 < JM-IRCRelay> [AlexCato1] yeah, i used -C sometimes if i really didnt care about privacy (like moving coins out of the maker-wallet, in which the coins were mixed hundreds of times anyways). But -N 0 works for use cases like that now
03:46 < JM-IRCRelay> [AlexCato1] cant see a reason what it's good for now
04:44 -!- rdymac [uid31665@gateway/web/irccloud.com/x-fxdugzizveayzhtq] has quit [Quit: Connection closed for inactivity]     
05:40 -!- rdymac [uid31665@gateway/web/irccloud.com/x-rchdlclkuioupuvt] has joined #joinmarket     
06:08 -!- Giszmo [~leo@pc-40-227-45-190.cm.vtr.net] has joined #joinmarket     
07:38 -!- btcdrak [uid165369@gateway/web/irccloud.com/x-wigihvkwbcatlrdi] has quit [Quit: Connection closed for inactivity]     
07:44 -!- rdymac [uid31665@gateway/web/irccloud.com/x-rchdlclkuioupuvt] has quit [Quit: Connection closed for inactivity]     
08:26 -!- btcdrak [uid165369@gateway/web/irccloud.com/x-djndphyymhrasttf] has joined #joinmarket     
10:23 -!- plorph [2fb9c2a0@gateway/web/freenode/ip.47.185.194.160] has joined #joinmarket     
10:48 -!- BonyM1 [~BonyM-I@ua-83-227-211-4.cust.bredbandsbolaget.se] has quit [Ping timeout: 265 seconds]     
11:01 -!- BonyM1 [~BonyM-I@ua-83-227-211-4.cust.bredbandsbolaget.se] has joined #joinmarket     
12:10 -!- plorph [2fb9c2a0@gateway/web/freenode/ip.47.185.194.160] has quit [Quit: Page closed]     
12:26 -!- arubi_ [~ese168@bzq-109-65-85-49.red.bezeqint.net] has joined #joinmarket     
12:27 -!- arubi_ [~ese168@bzq-109-65-85-49.red.bezeqint.net] has left #joinmarket []     
12:58 -!- viasil [~viasil@198.12.75.25] has quit [Ping timeout: 252 seconds]     
13:15 < belcher> agreed with removing -C
13:17 < belcher> im unbelievably irritated by realising the query-utxo-set thing, that kills any possible good lightweight sync'ing method
13:18 <@waxwing> belcher: i'm clearly not quite up to speed, what's particularly irritating? wouldn't we always need to query the utxo set?
13:18 < belcher> for those not following, the taker must be able to check that the maker's utxos are really unspent, existing spv methods arent able to provide this functionality, so the only way is either unprivate methods (blockr.io, electrum servers) or running a full node
13:18 < belcher> "for those not following" i wrote before you replied, though now it looks rather rude :\
13:20 < gmaxwell> One can just fail to check and all you lose is the protection for yourself. There really is no escape there, without the ability to check the utxo set you can be DOSed with invalid joins of totally non-existing coins.
13:20 < belcher> yes, its the DOS possibility that non-checking provides thats the problem
13:20 < belcher> you dont even know which maker was the culprit
13:21 < belcher> i guess you could ban all of them..
13:21 <@waxwing> there's room for discussion there; we are open to DOS of that *general type* anyway.
13:21 <@waxwing> "more than you can count" :)
13:21 <@waxwing> the not-able-to-tell-which is a good point though, that is worse for sure
13:22 < belcher> and also failing a join uses up one of your podle commitments
13:23 < belcher> i guess on balance it might be worth writing anyway, and if that attack ever happens then users will have to switch to using full nodes (the makers dont know for sure that the taker isnt checking)
13:26 < gmaxwell> pruned full nodes... or someone would need to stand up a service that provides private UTXO set accesses.
13:29 < belcher> i mentioned in the issue about this how a p2p version of estimatefee was mentioned as something after bip151, after that a helpful method would be a p2p version of gettxout, which would let takers using the "connect to a single node you trust" method 
13:30 < belcher> also petertodd's merkle mountain range ideas for the utxo set would work here too
13:36 -!- arubi [~ese168@unaffiliated/arubi] has quit [Ping timeout: 260 seconds]     
13:39 -!- arubi [~ese168@unaffiliated/arubi] has joined #joinmarket     
13:46 -!- arubi [~ese168@unaffiliated/arubi] has quit [Quit: Leaving]     
13:47 -!- arubi [~ese168@unaffiliated/arubi] has joined #joinmarket     
14:33 -!- viasil [~viasil@198.12.75.25] has joined #joinmarket     
14:49 < JM-IRCRelay> [AlexCato] found edge case bug (minor annoyance), will post about it tomorrow. In short: if a maker has a higher TX fee contribution than the final TX will cost, then this same maker will throw an error to the taker because the final cj amount does not match the expected value
14:49 < JM-IRCRelay> [AlexCato] workaround (maker side): dont add too much tx fee
14:50 < JM-IRCRelay> [AlexCato] on the taker side: the maker just gets marked as unresponsive after he threw the error, taker selects another one then
14:52 < JM-IRCRelay> [belcher] interesting, how did you find that bug?
14:53 < JM-IRCRelay> [AlexCato] wondered why i usually had many joins on my makers, but nearly none today. Played around a little, couldnt find anything. Then tried to do a sendpayment just with myself as maker, which failed with
14:53 < JM-IRCRelay> [AlexCato] <<privmsg on CgAn: nick=something message=!error Wrong cj_amount. I expect XXXX
14:53 < JM-IRCRelay> [AlexCato] then i put on the sherlock hat.
14:54 < JM-IRCRelay> [belcher] hmm i had the same.. barely any joins today
14:54 < JM-IRCRelay> [belcher] ok
14:55 < JM-IRCRelay> [AlexCato] the bug cant be the reason for the few joins though, as there needs to be "perfect storm" conditions for it to matter: few counterparties, high maker tx contribution, very low priority fee by taker
14:55 < JM-IRCRelay> [AlexCato] guess there's just not much traffic today
14:57 < JM-IRCRelay> [belcher] have you noticed much correlation with the bitcoin price? i remember a massive amount of joins around november 3rd when the price was rallying
14:59 < JM-IRCRelay> [AlexCato] had two huge joins yesterday, but the samplesize is just too low for any valid conclusions; my feeling says it looks rather coincidental, as i've seen consistent joins within the last weeks, but nearly none today (with big bitcoin price jumps up)
15:00 < JM-IRCRelay> [AlexCato] with the history function, it should be easily plottable though (btc price change vs. number of joins that day)
15:00 < JM-IRCRelay> [AlexCato] im not curious enough to do the work for that though :)
15:01 < JM-IRCRelay> [AlexCato] (afk)
15:58 -!- btcdrak [uid165369@gateway/web/irccloud.com/x-djndphyymhrasttf] has quit [Quit: Connection closed for inactivity]     
17:50 -!- coins123_ [~coins123@ip-244-225.sn1.clouditalia.com] has joined #joinmarket     
17:50 -!- coins123 [~coins123@unaffiliated/coins123] has quit [Read error: Connection reset by peer]     
17:55 -!- coins123 [~coins123@ip-244-225.sn1.clouditalia.com] has joined #joinmarket     
17:55 -!- coins123_ [~coins123@ip-244-225.sn1.clouditalia.com] has quit [Read error: Connection reset by peer]     
17:55 -!- coins123 [~coins123@ip-244-225.sn1.clouditalia.com] has quit [Changing host]     
17:55 -!- coins123 [~coins123@unaffiliated/coins123] has joined #joinmarket     
19:00 -!- Giszmo [~leo@pc-40-227-45-190.cm.vtr.net] has quit [Quit: Leaving.]     
19:16 -!- adlai [~adlai@unaffiliated/adlai] has quit [Ping timeout: 258 seconds]     
19:25 -!- adlai [~adlai@unaffiliated/adlai] has joined #joinmarket     
19:28 -!- btcdrak [uid165369@gateway/web/irccloud.com/x-xsecnlwrykzdqdwh] has joined #joinmarket     
22:01 -!- coins123 [~coins123@unaffiliated/coins123] has quit [Ping timeout: 240 seconds]     
22:17 -!- oman9978 [682041c9@gateway/web/freenode/ip.104.32.65.201] has joined #joinmarket     
22:26 -!- oman9978 [682041c9@gateway/web/freenode/ip.104.32.65.201] has quit [Quit: Page closed]     
23:23 -!- coins123 [~coins123@unaffiliated/coins123] has joined #joinmarket