--- Day changed Thu Sep 22 2016 00:30 < adlai> of course it's not recognizing any tx as a jmtx >_< 00:32 < adlai> it recognizes if i feed it a specific txid 00:37 < adlai> aha, pebkac - specifying only the start of a range does not, as i ASS-U-MEd, scan from that block to the latest one... 00:37 < adlai> it scans only that block. 00:40 * adlai launches it on a couple hundred blocks, watches mr robot now :D 00:45 < adlai> aha, fungibit forgot to mention "only run this on computers with MOST_POSITIVE_FIXNUM yottabytes of RAM, as it keeps EVERYTHING in memory" 00:50 * adlai runs it on 50 blocks; the progress bar is amusing, reading almost 3 hours but dropping by a minute every ~5 seconds 00:53 < adlai> bleh, now it's over 3hrs... kinda https://xkcd.com/612/ ish 00:54 < adlai> the inability to make an accurate estimation is understandable, since you can't know in advance how difficult the knext knapsack will be 01:53 -!- midnightmagic [~midnightm@unaffiliated/midnightmagic] has joined #joinmarket 02:41 < adlai> thoughts on https://www.reddit.com/r/joinmarket/comments/53qirp/btc_decentralized_peer_to_peer_tumbling/d7x933l ? 02:41 * adlai is a little worried OP has never touched a CLI before 02:45 < adlai> in other news, fungibit's unmixer keeps racing into an indefinite lockspin when run with -w 3 02:45 <@waxwing> adlai: interesting thoughts, thanks 02:46 <@waxwing> if i'm reading this right my bot has done at least 10-15 txs without reannouncing. was musing earlier today, maybe we need both extremes: bots that announce all the time so "real" reannouncements are lost in the mix, and "quiet" ones like mine. 02:47 <@waxwing> the problem with the noise adding approach is it can end up with outlandish data transfer eventually 02:47 < adlai> by outlandish you mean excessive? 02:47 <@waxwing> yeah :) 02:48 * adlai has always imagined that the Sufficiently Decentralized Channel will have a cost-to-broadcast (either like the hp2 commitments, or PoW, or an actual monetary cost (eg op_return fees)), so the invisible hand will tame that 02:49 <@waxwing> heh you're sneakily advocating for market announcements on chain, aren't you :) 02:50 <@waxwing> but yeah i can't see any counter-argument, at least in theory. 02:50 < adlai> pit advertisements is probably more realistic; makers still have an interest to allow other makers to participate in a pit that they host, because they could provide more liquidity that way 02:51 < adlai> so op_return pit ads could contain an IP/onion + port, and maybe a bit of info about the liquidity available therein 02:52 <@waxwing> yeah that's fine, but i'm not sure that kind of thing needs the blockchain does it. 02:52 < adlai> makes the pit's location difficult to censor 02:52 < adlai> (that would require miner collusion) 02:53 <@waxwing> yes but that's the kind of information that'd be near impossible to censor anyway 02:54 < adlai> the main problem i see with this approach is that makers need to take the initiative to host&advertise their own pits, and i'm not sure mutual distrust is strong enough to make them eat the upfront cost 02:54 < adlai> yeah. if somebody can find a copy of the JM code, they can probably also find a reference to a working pit 02:56 <@waxwing> as for having multiple pits, let alone advertisements it would be nice to just start running with more than 1 :) 02:56 <@waxwing> thanks for looking into fungibit's work more, it's great if we can get more solid understanding of that stuff. 02:57 <@waxwing> i guess you're better positioned than anyone else to do it too 02:57 < adlai> my guess is that incentives push towards a single pit (which could be a decentralized channel, but in the IRC world... isn't) 02:57 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Ping timeout: 244 seconds] 02:57 < adlai> https://joinmarket.me/obtest/ is down 02:58 <@waxwing> well a couple of things: i don't agree that *duplicated* pits are disincentivized, it works fine if more than 1 is default 02:58 <@waxwing> re: obtest yeah i'm being lazy. when my VPS cycled i didn't bring it back up sorry. 02:59 < adlai> piggybacking on existing IRC servers is "free", but not too resilient 02:59 <@waxwing> second point, re IRC/decentralized, i played around with a hidden service model, the problem is you might effectively rely on seed nodes to give you honest broadcast, because the long term participants have no motivation to serve you the full feed. 03:02 < adlai> in my imaginary model, takers would scan some large chunk of the recent chain looking for advertisements; then well capitalized makers (including wannabe snoops!) would advertise more often... 03:03 <@waxwing> i think offers kinda have to be real-time. 03:03 <@waxwing> maybe you can work around that, but it's a bit ugh. 03:03 < adlai> right, that's why i'm imagining pit offers rather than offer-offers 03:03 <@waxwing> right, and my feeling was that the blockchain properties don't seem to be needed for that. 03:04 <@waxwing> maybe there's something there, some hybrid thing. i can't quite envision it yet. 03:04 < adlai> we should just make this whole idea irrelevant by prioritizing the switch to matrix (or another channel, although iiuc this one works already?) 03:05 < adlai> it doesn't require a protocol upgrade either, the switch could theoretically happen gradually, starting yesterday 03:05 <@waxwing> i had a terrible time getting matrix working. it's very state-y. maybe someone who has more skill in it could get it working in a streamlined way. 03:05 <@waxwing> it did work, but there were delays connecting counterparties and what not. 03:06 <@waxwing> adlai: from what i saw, i see little advantage of matrix over IRC 03:06 <@waxwing> more important to have more than 1 pit, than to have different types of pit. 03:06 <@waxwing> i did a pit over websockets as well, that worked well, but just with a webserver 03:07 < adlai> ideally we'd have one pit, but it'd be an actual decentralized channel; i thought matrix worked that way. 03:07 <@waxwing> well the servers are federated, but it's not that different from IRC afaict. 03:07 <@waxwing> if you own a matrix homeserver you can still censor on it. 03:08 <@waxwing> maybe it's discoverable though, there seemed to be some discussion about that. but i never quite got that far, i didn't get it practically working in a satisfying way to begin with. 03:08 < adlai> https://github.com/AdamISZ/joinmarket/compare/matrixmc...AdamISZ:matrix-mc >> Showing 30 changed files with 1,123 additions and 137 deletions >_< 03:09 <@waxwing> it would just be one file now, if it was usable. 03:09 <@waxwing> but you need the backend code of course, what was it matrix-appserver something something 03:10 <@waxwing> working with the webserver version i got it running with a tiny MessageChannel implementation 03:11 <@waxwing> http://pastebin.com/RvyNS1fP 03:11 < adlai> but that requires each maker to host a webserver, right? 03:11 <@waxwing> well no it would be like IRC, there would be an owner. 03:11 < adlai> oic 03:11 <@waxwing> clients would be web clients in that version 03:12 <@waxwing> ofc that code snippet was just a start, i had to add more than that to do it more properly 03:12 < adlai> good timing! it gets automatically deleted in 10 hours 03:12 < adlai> maybe stick your WIP in a branch? 03:13 <@waxwing> yeah i think it's in a separate repo actually, but good point, i'll do that. but don't worry it won't disappear :) 03:14 <@waxwing> just to illustrate that there's an interface so that whatever weird stuff we do it should just plug into MessageChannel 03:22 <@waxwing> you could probably do an ultra simplistic pit server without anything flashy, use whatever infrastructure you like. have it serve TLS i guess is enough, with some rate limiting. 03:22 <@waxwing> you'd end up re-inventing the wheel though no doubt, with inter-user messaging being required. 03:23 <@waxwing> maybe some boilerplate code from python twisted :) 03:23 < adlai> i believe this is called miniircd :) 03:24 <@waxwing> right. btw what about Tor connections with that? 03:25 <@waxwing> i guess ordinary Tor it doesn't matter, right? and for HS maybe it also doesn't matter? 03:26 < adlai> didn't we test this? 03:27 < adlai> although not too intensely 03:27 <@waxwing> I don't know, but if we did it can address https://github.com/JoinMarket-Org/joinmarket/issues/615 03:27 <@waxwing> if we stick it into the test suite somehow? 03:28 < adlai> pull tor into our travis setup just to test socks.py? 03:28 <@waxwing> oh that was the problem, right. that's why i said we can't in that comment :) 03:29 <@waxwing> locally though? 03:29 <@waxwing> although maybe pulling it into travis is possible, haven't looked into it 03:30 < adlai> it should be entirely possible, i think we can pull any APT package 03:32 <@waxwing> sure 03:32 <@waxwing> we need some test code first 03:33 <@waxwing> even for things where it's not possible to run "real" one can mock up tests no doubt. 03:36 < adlai> lower-hanging fruit, what do you think of increasing the dust threshhold as suggested in https://www.reddit.com/r/joinmarket/comments/53c1lh/psa_how_to_avoid_tiny_outputs_into_the_yield/ 03:39 < adlai> raising this value does reduce maxsize, although it's an insignificant reduction compared to the vast majority of offers 03:50 <@waxwing> sorry i haven't thought about it carefully, my guess was it's good except that perhaps DUST_THRESHOLD should remain the actual blockchain threshold, but that's maybe an uninteresting point. apart from that i think i agreed with it. 03:55 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #joinmarket 03:57 < adlai> i don't think it should remain the blockchain threshhold. the btc dust threshhold is a "don't go below this" value, not a "go here and upwards", which is the effect of the jm dust threshold 03:58 <@waxwing> adlai: yes, all i meant is that the *variable* DUST_THRESHOLD should perhaps remain the real one. but again, not a very interesting point. 03:58 < adlai> i don't understand what you mean 03:58 <@waxwing> put it another way, "DUST_THRESHOLD" could reflect the one used in bitcoin proper, our own threshold could be a multiple of it. or something. 04:00 < adlai> mind if i dump this straight in develop? 04:00 <@waxwing> sure, if you think it doesn't need review 04:01 <@waxwing> btw if you do that it's good to run the tests locally first 04:02 <@waxwing> (we have yet another minor annoyance in that our tests being non-deterministic results in about a 1/2 chance of getting coverage failure (red cross). but that can stay on the back burner as far as i'm concerned.) 04:04 < adlai> what part of them is nondeterministic? 04:06 <@waxwing> some parts are easily addressed, like key generation (can use pseudo random). problem is, after that is done, there may still be non-determinisms from interaction with blockchain and network layer communications. 04:06 * adlai thought the tests used regtest? 04:08 <@waxwing> yes. the blockchain state will be different each time though, pretty sure. 04:44 -!- dserrano6 [~9dsd9s9da@dserrano5.es] has quit [Quit: upgrading] 04:45 -!- dserrano5 [~9dsd9s9da@dserrano5.es] has joined #joinmarket 05:10 < adlai> speaking of nondeterministic bugs, it looks as though fungibit's thing (i wish he'd given it a more unique name than jm_unmixer...) hangspins if bitcoind got a new block while it was running 05:10 * adlai has noticed correlation, still no idea why there would be causation here 05:11 <@waxwing> adlai: so he coded it to be usable with bitcoin running (i.e. not statically after?) 05:12 < adlai> it uses rpc to get blocks 05:12 < adlai> but it doesn't listen to blocknotify (shouldn't be hard for him to add that) 05:15 < adlai> ok, it got stuck this time without a new block 06:21 -!- twokun [~two@91.214.169.69] has joined #joinmarket 06:24 -!- fqtw__ [~me@x4db00312.dyn.telefonica.de] has quit [Ping timeout: 244 seconds] 06:32 -!- fqtw [~me@x4db00312.dyn.telefonica.de] has joined #joinmarket 06:41 -!- Giszmo [~leo@pc-40-227-45-190.cm.vtr.net] has joined #joinmarket 07:09 -!- twokun [~two@91.214.169.69] has quit [Ping timeout: 240 seconds] 07:20 -!- twokun [~two@91.214.169.69] has joined #joinmarket 07:24 -!- dddde33 [55386e26@gateway/web/freenode/ip.85.56.110.38] has joined #joinmarket 07:31 -!- twokun [~two@91.214.169.69] has quit [Remote host closed the connection] 07:36 -!- dddde33 [55386e26@gateway/web/freenode/ip.85.56.110.38] has quit [Quit: Page closed] 09:08 < GithubBot5678> [joinmarket] adlai pushed 1 new commit to develop: https://git.io/vi5NK 09:08 < GithubBot5678> joinmarket/develop e46cad9 Adlai: bump DUST_THRESHOLD to combat dust 09:31 -!- lnostdal [~lnostdal@90.149.74.80] has joined #joinmarket 09:58 <@waxwing> adlai: y u no sign? :p 10:00 -!- lnostdal [~lnostdal@90.149.74.80] has quit [Read error: Connection reset by peer] 12:06 -!- twokun [~two@46.19.137.116] has joined #joinmarket 12:35 -!- twokun [~two@46.19.137.116] has quit [Ping timeout: 244 seconds] 12:40 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has quit [Quit: ZNC - http://znc.sourceforge.net] 12:45 -!- luke-jr [~luke-jr@unaffiliated/luke-jr] has joined #joinmarket 12:46 -!- mustyoshi_ [~mustyoshi@2601:410:4303:4240:c9a0:6e14:3191:25d4] has joined #joinmarket 12:48 -!- mustyoshi [~mustyoshi@2601:410:4303:4240:d029:772:1c38:157d] has quit [Ping timeout: 248 seconds] 14:04 -!- rdbell [~rdbell@206.169.79.136] has joined #joinmarket 14:09 -!- megaddin [aladdin@gateway/shell/fnordserver.eu/x-zioulpjzjkhghmca] has quit [Quit: https://fnordserver.eu] 14:10 -!- twokun [~two@185.103.96.147] has joined #joinmarket 14:37 -!- Netsplit *.net <-> *.split quits: BigTrouble, RedEmerald, Emzy, core, zmanian__, OverlordQ, Lightsword, brand0, midnightmagic, coins123, (+4 more, use /NETSPLIT to show all of them) 14:37 -!- Netsplit *.net <-> *.split quits: BonyM1, rotarydialer, Pilate, berndj, q-biq, Einherjer 14:37 -!- Netsplit *.net <-> *.split quits: @waxwing, bertani 14:38 -!- Netsplit over, joins: pigeons, zmanian__, Lightsword, RedEmerald, brand0, twokun, BigTrouble, Emzy, core 14:39 -!- Netsplit over, joins: coins123 14:40 -!- Netsplit over, joins: @waxwing, berndj, rotarydialer, q-biq, bertani 14:40 -!- zmanian__ [sid113594@gateway/web/irccloud.com/x-otsenxlwizzbpiyc] has quit [Ping timeout: 252 seconds] 14:41 -!- Netsplit over, joins: midnightmagic, OverlordQ 14:41 -!- Netsplit over, joins: rdbell, puddinpop 14:42 -!- Einherjer [~einherjer@69.64.40.177] has joined #joinmarket 14:42 -!- Pilate [~pilate@infoforcefeed/pilate] has joined #joinmarket 14:42 -!- BonyM1 [~BonyM-I@ua-83-227-211-4.cust.bredbandsbolaget.se] has joined #joinmarket 15:16 -!- zmanian__ [sid113594@gateway/web/irccloud.com/x-oeopuvcnqstsenuo] has joined #joinmarket 15:51 -!- sgfdsfg34535 [55386e26@gateway/web/freenode/ip.85.56.110.38] has joined #joinmarket 15:53 -!- sgfdsfg34535 [55386e26@gateway/web/freenode/ip.85.56.110.38] has quit [Client Quit] 15:58 -!- megaddin [aladdin@gateway/shell/fnordserver.eu/x-znuuxgmjlbutanqq] has joined #joinmarket 17:10 -!- rdbell [~rdbell@206.169.79.136] has quit [Read error: Connection reset by peer] 19:30 -!- fqtw_ [~me@x590cd9b7.dyn.telefonica.de] has joined #joinmarket 19:33 -!- fqtw [~me@x4db00312.dyn.telefonica.de] has quit [Ping timeout: 244 seconds] 22:12 -!- Giszmo [~leo@pc-40-227-45-190.cm.vtr.net] has quit [Quit: Leaving.]