--- Day changed Sat Jan 07 2017
00:38 -!- lnostdal [~lnostdal@62.90-149-73.nextgentel.com] has quit [Ping timeout: 258 seconds]     
00:42 -!- lnostdal [~lnostdal@62.90-149-73.nextgentel.com] has joined #joinmarket     
01:29 -!- windsok [~windsok@45.63.59.8] has quit [Ping timeout: 256 seconds]     
01:57 -!- windsok [~windsok@45.63.59.8] has joined #joinmarket     
03:08 -!- joinmarket_first [~espen@160.121.9.46.customer.cdi.no] has joined #joinmarket     
03:08 < joinmarket_first> quick question: blockr.io is reporting my joinmarket address as a false address:
03:09 < joinmarket_first> http://blockr.io/address/info/12iwWRcSCUXDgVmc8erYCi1ghczzhfjhxH
03:09 < joinmarket_first> but according to blockchain.info, it should have a ballance.
03:10 < joinmarket_first> did I do something wrong when generating my joinmarket wallet?
03:10 < waxwing> joinmarket_first: look here: http://btc.blockr.io/
03:10 < waxwing> not too promising :(
03:10 < waxwing> last blocks: None
03:11 < waxwing> if you are using blockr, you will have to switch to using Core. blockr appears not to be working from what i can see. although i haven't tried the API.
03:11 < waxwing> in other words, i haven't tested the part of the website that joinmarket uses, right now, to ensure that i'm right in saying it isn't working. it just probably isn't working.
03:14 < JM-IRCRelay> [AlexCato1] ah, that might be the reason why there's so incredibly few coinjoins lately... blockr not working
03:15 < JM-IRCRelay> [AlexCato1] @joinmarket_first: nope, you didnt do anything wrong, i dont think there actually *IS* anything that you could do wrong in generating a JM wallet
03:41 < joinmarket_first> imported the private key to a wallet and was able to transfer the funds, so something is wrong with blockr.io. I was pretty nervous when it wasn't showing up and afraid I'd sent to a burn address.
03:41 < joinmarket_first> will try running full node.
07:54 < fluffypony> AlexCato1: was great meeting you too!
07:59 < grubles> >[AlexCato1] ah, that might be the reason why there's so incredibly few coinjoins lately... blockr not working <- ah i didn't think of that
07:59  * grubles shakes fist at blockr!
10:25 -!- so [~so@unaffiliated/so] has joined #joinmarket     
11:09 -!- windsok [~windsok@45.63.59.8] has quit [Ping timeout: 240 seconds]     
11:48 < waxwing> just tried a sync of my main wallet, my best guess from comparison is that it's serving the chain from a day or so behind (my joins in the last day are missing).
11:48 < waxwing> tbh that's actually much worse than it not working at all.
12:01 < lebeev> waxwing: I don't think twisted automatically fixes the "code is a mess" issue
12:01 < lebeev> but the twisted project on github does look much better
12:02 < waxwing> lebeev: well sure (as per above discussions), which project do you refer to?
12:02 < lebeev> joinmarket-clientserver
12:03 < lebeev> of course twisted isn't really the optimal solution for producing the cleanest possible code
12:03 < waxwing> ok, yes, that's my one. slight ambiguity because ghtdak did one before, but it's old now.
12:04 < lebeev> feels more coherent so far than the main repo
12:04 < waxwing> i think the whole asynchronous approach is better, easier to avoid un-understandable conditions or unrepeatable errors.
12:04 < lebeev> I think some design docs would go a long way
12:04 < waxwing> but i'd like any input
12:04 < lebeev> and perhaps some kind of a style guide
12:05 < waxwing> there is a joinmarket-docs repo in github.com/joinmarket-org
12:05 < waxwing> it's mainly to help anyone understand joinmarket's algorithms in more detail
12:05 < lebeev> im thinking more about the application design
12:07 < lebeev> helps a lot with maintaining coherence
12:08 < waxwing> for clarification, it's much more a refactoring than a rewrite, most of the core chunks of code are copy-pastes from the original. mostly. it's not a 100% rewrite because it's basically implementing the same functions.
12:08 < lebeev> yeah I know
12:09 < waxwing> as for design difference take as a starting point the intro. paragraph in the -clientserver readme
12:09 < lebeev> I've looked at it
12:09 < waxwing> but again, input most welcome.
12:09 < waxwing> hmm blockr testnet syncing worked fine, but i guess that's no surprise because they're old coins...
12:12 < waxwing> might be worth reminding people that the electrum and bc.i interface PRs are still there. both were waiting for author input as of last summer :(
12:12 < waxwing> yeah the new hybrid SPV mode of core does seem like a great idea right now :)
12:13 < lebeev> you taking pull requests?
12:13 < waxwing> absolutely
12:14 < lebeev> I've got some twisted devs working for me, and would rather like to replace the current joinmarket tumbler I use with something a little more stable
12:14 < lebeev> I'll try to get someone to take a proper look at it within the next few days
12:15 < waxwing> yeah ok. for that, i'm nearly there i *think*, because using the schedule file i can now generate tumbler schedules (tried it in Qt, it's working)
12:15 -!- HugsNotDrugs [~HugsNotDr@52.24.33.104] has joined #joinmarket     
12:15 < waxwing> and i think it's fairly small change to have it re-start using the current state of the wallet to generate the new schedule
12:16 < waxwing> the instability is hard to remove completely i think, but it should certainly possible for it to be restartable automatically.
12:19 < lebeev> being able to recover from issues would probably significantly increase the % of succesful tumbler runs
12:21 < waxwing> yeah we were discussing this the other day, currently there is recovery code that times out and restarts in threads (current code), i'm not sure but i have a sense that a "restartable" approach as mentioned above (dynamically re-generate schedule) might be better. but i haven't thought this through, so, we'll see.
12:23 -!- windsok [~windsok@45.63.59.8] has joined #joinmarket     
12:39 -!- btcdrak [uid165369@gateway/web/irccloud.com/x-vfjscqeglycdzyon] has quit [Quit: Connection closed for inactivity]     
12:57 < waxwing> if anyone feels like testing the -Qt binary here: https://github.com/AdamISZ/joinmarket-clientserver/releases/ that'd be helpful; i'd expect bugs though so don't use it for real (except for testing loading/syncing wallets if you like).
13:03 -!- xpub [253074ba@gateway/web/freenode/ip.37.48.116.186] has joined #joinmarket     
13:09 < belcher> ill try it
13:10 < waxwing> thanks belcher . i'm sitting here twiddling my thumbs trying to decide if it's worth setting up some testnet ygs again. because testing over blockr is now a fail.
13:10 < waxwing> of course, more to the point, before something like the core spv thing, JM appears no longer to be usable without a full node.
13:11 < belcher> waxwing i got segmentation fault
13:11 < belcher> any idea how to get debug information?
13:11 < waxwing> btw that Qt thing can be tested on regtest (that's mostly what i do of course).
13:11 < waxwing> belcher: not sure, which flavour of OS is it?
13:11 < belcher> xubuntu
13:12 < waxwing> i just tested on a clean ubuntu 1404 vm
13:12 < belcher> you ran with ./joinmarket-qt ?
13:12 < waxwing> yeah
13:12 < belcher> with chmod +x joinmarket-qt beforehand
13:13 < waxwing> it shouldn't need anything else, that's why i tested from clean. presumably some dynamic library mismatch type failure, dunno.
13:14 < waxwing> pretty sure it's not something stupid like 'it expects the directory wallets/ to be there', that's why i tested it from clean.
13:14 < waxwing> oh i just kind of repeated myself there sorry :)
13:15 < belcher> i just ran it in gdb
13:15 < belcher> Program received signal SIGSEGV, Segmentation fault.
13:15 < belcher> strlen () at ../sysdeps/x86_64/strlen.S:106
13:15 < belcher> 106	../sysdeps/x86_64/strlen.S: No such file or directory.
13:15 < belcher> somewhere in strlen then..?
13:16 < waxwing> probably a C library mismatch eh? not sure. in the past i found binaries built on ubuntu were not usable on debian iirc.
13:16 -!- btcdrak [uid165369@gateway/web/irccloud.com/x-suxnsdefkfbqizoo] has joined #joinmarket     
13:16 < belcher> are you able to make a .exe ?
13:17 < waxwing> haven't tried yet.
13:17 < belcher> more people use windows after all, and i get a feeling linux users are happier to run lots of sudo apt-get install lines that make something work
13:17 < waxwing> was going to go for TAILS and windows later. it all seems a bit academic if people can't use blockr atm and have no other choices.
13:17 < waxwing> well, not so much for windows i guess.
13:17 < belcher> ah tails yes
13:18 < waxwing> yes, i undertand that, it was more a matter of starting with the easiest case and going from there.
13:18 < waxwing> in theory i'm re-following an old path (JMBinary). there i had a 'debian32', 'ubuntu64' and 'exe' (windows). but at that time 'debian32' was what was needed for TAILS. TAILS has moved on in terms of its arch.
13:18 < belcher> mmm     
13:20 < waxwing> binary creation is just pyinstaller.py -F joinmarket-qt.py ; that simple, in principle.
13:23 < waxwing> which version of xubuntu belcher ?
13:24 < belcher> Description:	Ubuntu 16.04.1 LTS
13:24 < belcher> Codename:	xenial
13:25 < waxwing> right, i'm downloading it now. although i might be better off testing ubuntu 1604. oh well, let's see.
13:25 < belcher> so you were testing on 16.04.0 ?
13:26 < waxwing> no i tested on a 1404 vm.
13:26 < belcher> ah     
13:26 < waxwing> i run 1510 for some reason :)
13:37 < waxwing> oh well replicated the segfault ok :)
13:37 < belcher> :)     
14:09 < waxwing> confirmed on ubuntu 1604 too. seems to be related to this: https://github.com/pyinstaller/pyinstaller/issues/2232
14:15 < belcher> lots of discussion there
14:16 < waxwing> yeah sorry it's in the weeds and rather off topic; see this https://github.com/pyinstaller/pyinstaller/issues/2232#issuecomment-260352159 and previous comment, it's very clear that this is the issue.
14:16 < waxwing> i did 'bt' in gdb and saw that LD_LIBRARY_PATH and env stuff too.
14:32 < waxwing> belcher: fixed, can you try again? should work, works on my xubuntu vm
14:32 < waxwing> no rush i'm off to bed soon anyway
14:34 < belcher> sec     
14:39 < belcher> looks like it works waxwing
14:39 < belcher> the gui comes up and everything
14:39 < belcher> what was the fix btw?
14:39 < waxwing> they just had a bug in their binaries. as per the issue.
14:40 < belcher> right so you just reverted to before that commit.
14:40 < waxwing> well actually i went forwards to the latest; both were viable options
14:40 < belcher> ah lol
14:40 < waxwing> but like it said there you had to rebuild the binaries, it doesn't update automatically. a bit in the weeds but anyhoo.
14:41 < waxwing> this seems common, i had a similar nasty bug the first time i tried this (different bug but similarly a case where i had to "move around" in the repo)
14:59 < adlai> waxwing: adlai: i quite like alphabetical << how does one type the sound made by snorting air through the sinuses while awake
15:02 < adlai> quitobro: you guys are able to use Bitcoin Unlimited for lower fees, is that what's going on? << i'll have what he's having!
15:03 < adlai> the sheer beauty of jm NOT being consensus software is that it benefits from having multiple implementations yesterday
15:03 < adlai> btc... tmrw
15:06 -!- juscamarena__ [~justin@47.148.176.74] has joined #joinmarket     
15:49 -!- windsok [~windsok@45.63.59.8] has quit [Ping timeout: 240 seconds]     
16:08 -!- juscamarena [~justin@47.148.176.74] has joined #joinmarket     
16:24 -!- windsok [~windsok@45.63.59.8] has joined #joinmarket     
16:47 < JM-IRCRelay> [AlexCato1] blockr.io not working: havent thought about it much, just throwing it out here: what's the effect on running yield generators, which use blockr? This could cause a lot of unintentinal self-DoS, as probably no taker commitments will be accepted (cause the YG cant see the UTXOs). Maybe theres few coinjoins because the commitments just get used up due to
16:47 < JM-IRCRelay> a certain amount of YGs not accepting them
17:04 < lebeev> that could explain the super high tumbler failure rates
17:14 -!- GAit [~GAit@unaffiliated/gait] has quit [Read error: Connection reset by peer]     
17:27 -!- GAit [~GAit@unaffiliated/gait] has joined #joinmarket     
17:55 < belcher> AlexCato1: the vast majority of yieldgens use bitcoin core, only 3 right now that use blockr.io
17:56 < belcher> yieldgens put what blockchaininterface they use in their irc realname so that we can monitor the situation of whether they use core or not, use /who #joinmarket-pit to see them all
18:04 < JM-IRCRelay> [AlexCato1] ah, a feature i didnt know about. Great forward thinking!
18:05 < gmaxwell> not great for privacy though, esp for the blockr users. :P
18:06 < belcher> if a bot is anything but a yieldgen, that field wont say anything
18:09 < belcher> the realname isnt used for anything in the code, its only to see how many yieldgens follow the advice from the tutorial, short term thinking yieldgens have no incentive to use core over blockr so it was always a question in my head of how many would actually do so
18:11 < grubles> heh good to know :)
18:11 < belcher> controversial opinion incoming: im glad blockr is down, that code never shouldve been written
18:12 < belcher> though i guess it comes down to the same old debate of paranoia and security vs usability and practicality
18:12 < JM-IRCRelay> [AlexCato1] well, I prefer the good solution thats actually used over the perfect one that isnt
18:12 < grubles> forgive my ignorance, but can jm use existing electrum servers instead of blockr type of sites?
18:13 < JM-IRCRelay> [AlexCato1] this is probably especially hard on Tails users, which cant just set up a full node
18:13 < belcher> grubles theres a pull request for it, https://github.com/JoinMarket-Org/joinmarket/pull/475
18:14 < JM-IRCRelay> [AlexCato1] (fully on your side for yield gens: they should use full nodes. Takers on the other side...)
18:14 < belcher> im cautiously excited about this idea: https://github.com/JoinMarket-Org/joinmarket/issues/653 where bots connect to the bitcoin p2p network and download full blocks starting from their wallet creation date and parse them
18:15 < belcher> it has SPV security (so trusts the miners) but full privacy like a node, and has no setup so users could just turn it on and use it without configuring anything
18:15 < belcher> full privacy like a node except for leaking your wallet creation date*
18:16 < belcher> if you create a wallet and just start using it, you dont even need to download historical blocks since the wallet creation date is now
18:17 < JM-IRCRelay> [AlexCato1] awesome, that should be a nice default setting for a fresh JM
18:19 < gmaxwell> you need a utxo set too, no, for verifying that coins being spent exist?
18:19 < belcher> yeah thats an issue, right now im thinking just trust the yield generators... worst that can happen is the coinjoin is actually invalid
18:20 < belcher> yieldgens dont get paid if they send invalid utxos after all, so maybe it will work
18:22 < JM-IRCRelay> [AlexCato1] thats another DoS vector though. Have a few cheap YGs, make the takers wonder why their (invalid, but they dont know it) coinjoins dont confirm?
18:24 < belcher> other nodes on the p2p network would tell them their transactions are invalid, after that the taker can exclude all the yieldgens that took part from future coinjoins and try to recover