--- Day changed Thu Oct 20 2016 00:18 < waxwing> nkuttler: do you need coins for your faucet? or which one is the most deserving/needy :) 00:19 -!- devn0ll [~devnull@qwerty.cloud.tilaa.com] has left #joinmarket [] 00:35 < waxwing> life is a lot easier with the direct send feature :) moving all kinds of testnet coins around here :) 00:35 < waxwing> btw there were like 300 testnet coins in "wallets/wallet.json" and "wallet.json" (interpreted as seed because file was not there). quite amusing :) 00:52 -!- moli [~molly@unaffiliated/molly] has quit [Read error: Connection reset by peer] 00:52 -!- moli [~molly@unaffiliated/molly] has joined #joinmarket 01:17 -!- molz [~molly@unaffiliated/molly] has joined #joinmarket 01:21 -!- moli [~molly@unaffiliated/molly] has quit [Ping timeout: 260 seconds] 01:31 -!- solace [~h@wowana.me] has quit [Ping timeout: 250 seconds] 01:34 -!- solace [~h@wowana.me] has joined #joinmarket 01:38 < nkuttler> waxwing: i'm good, thanks 01:39 -!- solace [~h@wowana.me] has quit [Ping timeout: 260 seconds] 01:58 -!- solace [h@wowana.me] has joined #joinmarket 03:41 < waxwing> belcher: thanks for feedback. my memory is that it's bad that the any of the commitment functionality is in podle.py, that should have been purely the ECC stuff, put off any refactor till later. some other uninteresting details that can be twiddled later. 03:41 < waxwing> the issue du jour is, what is a reasonable way to dissuade this duplicated bot (or others) that doesn't create side-effects. 03:49 < nkuttler> btw, can jm use blockr over tor? 04:00 < waxwing> nkuttler: yes, but it doesn't avoid the biggest privacy issue right, ie requesting utxos in groups is strongly suggestive of common ownership 04:17 < GithubBot5678> [joinmarket] AdamISZ closed pull request #642: add hostid to debug log (develop...patch-19) https://git.io/vPHMi 04:47 -!- Cory [~Cory@unaffiliated/cory] has quit [Ping timeout: 245 seconds] 05:05 < waxwing> at the moment i am not seeing even a valid 'partial' solution except: utxos offered are signed for, *and* Makers accept multiple !tx messages (thus a taker could re-create a transaction with a new maker set without going back to the beginning and using another commitment) 05:06 < waxwing> it's clunky, needs a new offertype, and not a full solution 05:07 < waxwing> adlai: in other news, did you miss the DUST_THRESHOLD thing? 05:08 < waxwing> pm-ing a copy of the relevant bit 05:37 < waxwing> ok, consider the following: the current misbehaving bots do not respond to !auth, that is, they don't provide !ioauth with their utxos. they could do similar by not sending !pubkey in response to !fill. 05:37 < waxwing> This behaviour, at least, can be dealt with without any shenanigans: request N*2 !fills instead of N, then choose any of the N*2 that actually respond. 05:38 < waxwing> 2 is obviously arbitrary, see previous gist. 05:39 < waxwing> additional algos that make a smarter selection than randomly choosing N from N/2 I would like to put off, since unless we have verification of the utxos provided, we are opening cross-maker attacks through lying (e.g. banning conflicting utxos, or more subtle choices e.g. age) 05:51 -!- mkarrer [~mkarrer@7.red-83-47-85.dynamicip.rima-tde.net] has quit [] 05:53 -!- mkarrer [~mkarrer@7.red-83-47-85.dynamicip.rima-tde.net] has joined #joinmarket 06:09 < waxwing> adlai: i've just tried out making some edits in a branch to do the above ^ , now going to test it. 06:30 -!- lnostdal [~lnostdal@90.149.73.62] has quit [Read error: Connection reset by peer] 06:31 -!- lnostdal [~lnostdal@90.149.73.62] has joined #joinmarket 06:36 < waxwing> looks to be working, for sendpayment trials, without an excessive edit list. basically tweak it to respect (configured multiple)*N orders, pass N as well as the extended orderlist to CoinjoinTX, then break out of recv_txio loop when N is reached, reset nonrespondants to "actualrespondants", then the second phase proceeds as normal. 06:37 < waxwing> that is the dumbest conceivable way, i.e. just use the first N that complete the handshake. also favours faster response which isn't a bad thing. 06:38 < waxwing> heh, got a txid "txid = fee765a5...." , that was confusing for a second :) 07:24 -!- Giszmo [~leo@pc-40-227-45-190.cm.vtr.net] has joined #joinmarket 07:57 -!- Cory [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has joined #joinmarket 07:57 -!- Cory [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has quit [Excess Flood] 07:58 -!- Cory [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has joined #joinmarket 07:58 -!- Cory [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has quit [Excess Flood] 07:58 -!- Guest12765 [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has joined #joinmarket 07:58 -!- Guest12765 [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has quit [Excess Flood] 07:59 -!- Guest12765 [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has joined #joinmarket 10:52 -!- gmaxwell [greg@wikimedia/KatWalsh/x-0001] has quit [Read error: Connection reset by peer] 12:09 -!- grubles [~grubles@unaffiliated/grubles] has quit [Quit: Leaving] 12:13 -!- Guest12765 [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has quit [] 12:15 -!- Cory [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has joined #joinmarket 12:15 -!- Cory [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has quit [Excess Flood] 12:16 -!- grubles [~grubles@unaffiliated/grubles] has joined #joinmarket 12:19 -!- Cory [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has joined #joinmarket 12:19 -!- Cory [~Cory@24-240-67-80.dhcp.mdsn.wi.charter.com] has quit [Excess Flood] 12:25 -!- Cory [~Cory@unaffiliated/cory] has joined #joinmarket 13:14 < JM-IRCRelay> [AlexCato] (...) use the first N that complete the handshake. also favours faster response which isn't a bad thing. <-- i'd be unhappy with that, as it disincentivizes TOR usage 13:15 < JM-IRCRelay> [AlexCato] off-the-bat-idea: request the 2*N, get the offers, then randomly choose from them. That solves the problem 13:16 < JM-IRCRelay> [AlexCato] btw, thanks for the YG/script repo. I'll not have time to do anything with it until the coming weekend though :) 13:16 < waxwing> yah, it should end up something like that. and the Tor point is worth making, although i'm not 100% convinced either way 13:17 < waxwing> i found today that although a primitive test was easy, there are tricky difficulties when you grab N*2 orders, because you don't immediately fix the fee. 13:18 < JM-IRCRelay> [AlexCato] why would the fee change within a few seconds? 13:18 < waxwing> well you want to use N makers, but you haven't decided at the point of choose_orders exactly which, so you have a range of possible fees, but not the exact total yet. 13:19 < waxwing> it's an implementation detail, hopefully, at least for the non-sweep case, just needs more code 13:19 < waxwing> did you accept the invite then? 13:20 < JM-IRCRelay> [AlexCato] not yet, guess there's a mail waiting for me. Will check that out tomorrow as well 13:21 < waxwing> yeah there's some "collaborator" feature which allows, i think, to give people write access to specific repos. hopefully it works ok. 13:21 < JM-IRCRelay> [AlexCato] but i'll gladly help, so sure 13:22 < JM-IRCRelay> [AlexCato] and yeah, i see the problem with confirming the total fee, which is unknown at that point with the current way things works 13:22 < JM-IRCRelay> [AlexCato] *work 13:23 < JM-IRCRelay> [AlexCato] no immediate solution comes to mind; though there's at this very moment also an unknown: the miner fees vary wildly and arent displayed either 13:24 < JM-IRCRelay> [AlexCato] could display e.g. the average fee of the 2*N and give a range of the cj fee to be expected 13:24 < waxwing> alexcato btw, i'm off in a bit, but i believe the way you set the log level in the config doesn't allow overwrite from joinmarket.cfg 13:25 < JM-IRCRelay> [AlexCato] oh? Tested that. Will test again, though as well on the weekend 13:25 < waxwing> not 100% sure, but this happens before reading the file: https://github.com/JoinMarket-Org/joinmarket/blob/develop/joinmarket/configure.py#L308 13:26 < waxwing> well, seems fairly clear 13:28 < JM-IRCRelay> [AlexCato] interesting, and you must obviously be right. Wonder why I didnt catch that, because i did indeed change the value in the .cfg to test. In earlier versions the cfg-code was at the end of that method though, maybe i missed it when i put it at the top for reasons i cant remember 13:28 < JM-IRCRelay> [AlexCato] will fix that as well on the weekend, as i'll be gone for now as well. Sorry for that, good catch 13:28 < waxwing> yes, my first thought was 'put it at the end', but i bet you moved it to the start to make sure logging starts early or something 13:29 < waxwing> k, cya 13:29 < JM-IRCRelay> [AlexCato] possibly 13:29 < JM-IRCRelay> [AlexCato] ciao :) 21:36 -!- Giszmo [~leo@pc-40-227-45-190.cm.vtr.net] has quit [Quit: Leaving.] 22:06 -!- belcher [~belcher@unaffiliated/belcher] has quit [Ping timeout: 250 seconds] 22:16 -!- belcher [~belcher@unaffiliated/belcher] has joined #joinmarket