--- Day changed Fri Dec 28 2018 02:19 -!- asymptotically [~asymptoti@gateway/tor-sasl/asymptotically] has joined #joinmarket 06:10 -!- bsm117532 [~mcelrath@c-24-61-184-150.hsd1.ma.comcast.net] has joined #joinmarket 07:10 -!- bsm117532 [~mcelrath@c-24-61-184-150.hsd1.ma.comcast.net] has quit [Ping timeout: 250 seconds] 07:39 < waxwing> someone just succeeded in distracting me for an entire day fitting bitcoin price to a power law trend :) 07:40 < waxwing> was trying to replicate this old graph https://www.google.lt/imgres?imgurl=https%3A%2F%2Fsteemitimages.com%2F0x0%2Fhttps%3A%2F%2Fsteemitimages.com%2FDQmSVk9VYEmBuMxKjE29QXvzwa6ggGHY1amyG1cp98cGB3c%2Fip.bitcointalk.org.jpg&imgrefurl=https%3A%2F%2Fsteemkr.com%2Fbitcoin%2F%40adamantine%2Fbitcoin-logarithmic-regression&tbnid=wqE6gxc3444NmM&vet=1&docid=Ntm2_pjlb6R2ZM&w=1134&h=698&hl=en-lt 07:40 < waxwing> nice work google 07:40 < waxwing> probably like 4 viruses embedded in there, sorry 07:56 < arubi> lol 08:09 < qubenix> i made a tx two days ago with -N 0 and tx_fees = 1000 that still is unconfirmed. is that interesting at all? 08:09 < waxwing> qubenix, i don't know; what actual fee in satoshis shows up on the blockchain? 08:09 < waxwing> or on your node or w/e 08:09 < waxwing> not the blockchain, ofc :) 08:11 < qubenix> looks like fees were 124 and vsize was 134 08:11 < waxwing> 1 sat/b could be too small eh ... yeah, that number is ofc small, but i presume it got relayed? 08:12 < waxwing> iirc (can check) we always randomise the fee, iirc it's +/- 20%. 08:12 < qubenix> yeah, i had my relayfee turned down. some explorers show it, some dont 08:13 < waxwing> ok. yes i just checked, if you use the "static" (like 1000) option, we do a 20% above or below randomization. 08:13 < waxwing> guess we should warn if you end up < 1 sat/byte by old measure due to relay etc. 08:19 < waxwing> also we should have rbf .. has anyone actually thought about that in JM (I haven't). 08:20 < qubenix> so is there anything i can do besides wait? i was thinking set listunspent to 0 and cpfp it. will that work? it's not necessary for me to fix it, i'm just curious really. 08:22 < arubi> cpfp might help in getting it mined but won't help with making the parent relay (at least I don't think it would) 08:22 < arubi> for rbf you'll need a second round of signing, but I guess it's not a problem if all makers are online 08:23 < waxwing> arubi, it was -N 0 08:23 < waxwing> so just him 08:23 < waxwing> and it wasn't rbf because we don't have that :) 08:24 < waxwing> hmm, i'm curious, if you send something with v low fee and it gets only partial relay, and then you just double spend (no rbf), will it propagate and mine? 08:25 < waxwing> arubi, oh you were answering my 'has anyone thought about rbf' Q, sorry 08:25 < arubi> ah yea but I thought it was related :) didn't realize the -N 0 there 08:26 < arubi> also yes it would probably propagate to nodes that first rejected the too low fee 08:27 < arubi> I think that's how the simple double spends work 08:27 < qubenix> i've seen this happen before, but was always with makers and eventually always gets fixed with a double spend of one of the other makers inputs 08:32 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] so is there anything i can do besides wait? i was thinking set listunspent to 0 and cpfp it. will that work? it's not necessary for me to fix it, i'm just curious really. 08:32 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] yes, you can create a child transaction and spend the unconfirmed output with a fee that is high enough so both transactions are picked up 08:33 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] double spend works as well, but it's a pain to set this up in JM from my experience 08:35 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] you'll have to modify joinmarket.cfg, "listunspent_args" so you can spend an unconfirmed transaction (theres an example in the config how to do that). Worked for me in the past when I had that problem 08:37 < qubenix> AlexMaroney: so the fee on the child tx should be enough for that tx + enough to bump the parents fee or enough to pay the entire parent tx? 08:38 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] correct. So if you send another N=0 transaction using the unconfirmed funds as input, it will confirm both rather quickly if you choose 2 sat/byte for that new transaction 08:38 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] that brings the overall tx fee up enough to be picked up 08:39 < qubenix> thanks, i'll try it and see what happens for an experiment. 08:40 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] i've tried this in the past when there were unexpected fee spikes. Never had the issue to go below 1 sat/byte though. But if that method fails, you can still do the double spend that waxwing suggests afterwards 08:40 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] no funds at risk by trying 08:45 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] forgot to mention one thing: this method assumes that there's funds in your joinmarket-wallet to be spent from the low-fee transaction. Can be either the original spending amount (if the target address was one of your joinmarket addresses). Or you can spend the change amount of your original transaction. Both will work 08:45 < AgoraRelay> and it 08:45 < AgoraRelay> [agora-irc/CgRelayBot] will confirm both transactions (your original one and the new one) 08:47 < qubenix> right, it was a tx to my own wallet and it was a wallet i set up just to test something. so this was kind of the absolute best place to have this issue because it's completely isolated and really small amounts. 09:28 < waxwing> qubenix, sorry for the trouble. i'm curious why you chose 1000; i'm betting your thinking was 'i'll use the lowest allowable', but we let you down a bit because we haven't really publicised that it's 20% randomized (to prevent people watermarking all their JM txs). 09:28 < waxwing> i mean it's doubtless in some release notes, but i don't think even i could find it easily :) 09:32 < qubenix> yeah, i chose 1000 thinking i would end up with lowest allowable. 09:33 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] yeah, it should have been added to the joinmarket.cfg comment at least. I'll take care of that soon (AlexCato here btw, just temporary using other name for unrelated reasons). Would it also make sense to have a hardcoded floor of 1000 sat/kbyte? Currently the lowest possible value is 145 sat/kbyte 09:33 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] ("possible" as in "can be set in joinmarket") 12:16 < waxwing> belcher, arubi others, laurentmt has offered some very interesting input on the gist comments https://gist.github.com/AdamISZ/4551b947789d3216bacfcb7af25e029e 12:21 < arubi> > "IF amount(payment_alice) < amount(change_alice) THEN ..." maybe I'm not understanding the notation. how can alice's change be larger than her input? 12:26 < arubi> ohh, payment means output to bob. doh 12:27 < arubi> well not output, the amount alice sends bob. yea I had the wrong thing in mind 12:33 < waxwing> does anyone here know how bustabit's bustapay/p2ep thing works? 12:33 < waxwing> how does the user's wallet participate i wonder? 12:36 * waxwing stops being distracted and gets back to this wallet selection thing 12:44 < waxwing> arubi, i still don't get the if clauses (as i just noted), is there a typo or something, i think so 12:45 < waxwing> i got the general impression that "UIH2" means bob's contribution's going to have to be similar to alice's, which is interesting if true (not good, but interesting) 12:45 < arubi> the inequalities were flipped just now (inner ones) 12:46 < arubi> at least I think they did 12:47 < waxwing> arubi, the IF parts are flipped, so as to cover > and < ; but the things after the IF THEN are the same except an equals sign, no? 12:47 < arubi> I'm still reading the comments here 12:47 < waxwing> k, no rush, sorry :) 12:56 < AgoraRelay> [agora-irc/Hugus] Are there versions of joinmarket for other coins? 12:59 < arubi> hmm yea that's confusing :) so I think I agree about the first IF clause, but maybe for the second one it should be "amount(inputs_bob) >= amount(payment_alice)" ? 13:00 < arubi> that kinda transforms it to behave like the first IF, if we say even though it's obvious that this is a p2ep, the payer can't be known from that kind of tx 13:02 < arubi> well, p2ep or a payment to two destinations (although that would be weird with no change at all) 13:02 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/AlexMaroney] @Hugus: not that I'm aware of 13:04 < arubi> asked in the gist itself as well 13:21 < AgoraRelay> [agora-irc/Hugus] Thank you CgRelayBot 14:25 -!- GitHub18 [GitHub18@gateway/service/github.com/x-httrnvptmwdmoigf] has joined #joinmarket 14:25 < GitHub18> [joinmarket-clientserver] AlexCato opened pull request #271: Improve manually selected tx fee documentation (master...patch-1) https://git.io/fhIaq 14:25 -!- GitHub18 [GitHub18@gateway/service/github.com/x-httrnvptmwdmoigf] has left #joinmarket [] 16:07 -!- asymptotically [~asymptoti@gateway/tor-sasl/asymptotically] has quit [Quit: Leaving] 16:48 < belcher> waxwing i came up with the phrase "unnecessary input heuristic" myself, but the concept is much older, iv seen a video of a talk by jonas nick and christian decker from ~2015 where they mention the idea 16:49 < belcher> but the names given to the idea i found were slightly confusing and i think UIH is a better name so i used it when writing my not-yet-finished literature review of all bitcoin privacy ideas 16:50 < belcher> this UIH2 is very interesting, i completely missed it, it only applies to p2ep things and not to any regular transactions like UIH1 does 16:50 < belcher> the fact that UIH2 exists is an argument that the phrase "unnecessary input heuristic" is a bad phrase 16:56 -!- GitHub95 [GitHub95@gateway/service/github.com/x-vradshptcuihiwmw] has joined #joinmarket 16:56 < GitHub95> [joinmarket-clientserver] AlexCato opened pull request #272: Add tx fee floor if manually chosen (master...patch-3) https://git.io/fhIKt 16:56 -!- GitHub95 [GitHub95@gateway/service/github.com/x-vradshptcuihiwmw] has left #joinmarket [] 17:55 -!- AgoraRelay [~jmrelayfn@p5DE4AB64.dip0.t-ipconnect.de] has quit [Ping timeout: 246 seconds] 18:08 -!- AgoraRelay [~jmrelayfn@p54866B2E.dip0.t-ipconnect.de] has joined #joinmarket