--- Day changed Mon Oct 29 2018
02:03 -!- undeath [~undeath@hashcat/team/undeath] has joined #joinmarket     
04:17 -!- technonerd [~techno@unaffiliated/technonerd] has quit [Ping timeout: 260 seconds]     
04:31 -!- technonerd [~techno@unaffiliated/technonerd] has joined #joinmarket     
05:09 -!- lnostdal [~lnostdal@77.70.119.51] has quit [Ping timeout: 252 seconds]     
05:33 -!- azizLIGHT [~azizLIGHT@unaffiliated/azizlight] has quit [Ping timeout: 246 seconds]     
06:45 < waxwing> undeath, don't know if you're around today, but added several comments. i'm leaning towards user prompts for this feature.
06:45 < waxwing> belcher_, too: (see PR #166)
06:45 < belcher_> looking now
06:46 < AgoraRelay> [agora-irc/CgRelayBot] [cgan/belcher_] btw looks like the bridges work now?
06:46 < waxwing> yeah did -r yesterday, so that's nice, thanks for checking
06:46 < waxwing> oh i could have checked it myself, heh
06:48 < belcher_> user prompt is a good idea
06:48 < belcher_> "what is the maximum you're willing to pay for a coinjoin?"
06:48 < belcher_> maybe with some suggested options (because of the anchoring effect)
06:49 < waxwing> hmm i realise suddenly that i'm missing the point slightly: i'm thinking only of a sendpayment scenario, which is obv the least interesting one (unless they do it in a sequence).
06:49 < waxwing> i guess, the same thing works with tumbler though, since it's reusing maxcjfee
06:49 < waxwing> we can still prompt with guidance there, and (if they choose this option) the same will apply
06:50 < waxwing> it's slightly overkill ofc, the expert user will quickly tire of having to do this.
06:50 < belcher_> [ ] dont ask me again
06:50 < belcher_> ?     
06:50 < waxwing> hmm. yeah if it means they use the same one each time it's ok as long as they've chosen a custom value at least once.
06:51 < belcher_> or another way, have an option in the config file, max_cj_fee, which by default is `ask-me`
06:51 < waxwing> so i can buy that, yeah. unfortunately many users will choose the first anchored value available, and 'never ask again'.
06:51 < waxwing> and then we're back to square one :)
06:51 < belcher_> then an expert user can set `max_cj_fee = ` to what they want and it wont ask
06:51 < belcher_> hmm     
06:51 < belcher_> how about... the suggested value we show them is randomly chosen
06:52 < belcher_> (dont laugh lol)
06:52 < belcher_> or we could just not give any suggestions, and instead suggest to the user that they run ob-watcher.py to see
06:53 < belcher_> since thats how any other prices are used by people, they look at whats on offer then make a decision
06:53 < waxwing> the root of the problem is not user laziness, it's that it's genuinely difficult to understand what it means. even small gaps in knowledge leaves a user giving up and saying  'do it for me, i don't have time for this'.
06:53 < waxwing> yeah people just don't run ob-watcher. it's another thing to do and it annoys them.
06:54 < waxwing> your randomized suggested value isn't bad imo.
06:55 < belcher_> well thats how market mechanisms work, people have to make decisions about price at some point
06:55 < waxwing> i'd see it as like: brief paragraph: here's a (randomized in a range) suggestion: this is what it means: enter your own choice or "yes to default". and then "don't ask me again?" or whatever.
06:55 < waxwing> yes true, good point. but this was hidden in existing weighted order choose. it's just a binary choice.
06:55 < waxwing> agree to this price? y/m
06:56 < waxwing> it'd be the same here i guess;  you still get to reject if you like (well not in tumbler, though, hmm)
06:56 < waxwing> tumbler was already kind of working with this max cutoff function; difference was it was weighted, under that.
06:59 < waxwing> re: allowing a 'don't ask me again', perhaps the simplest is: if a config val is set: max-cj-fee = 0.01,10000, use that silently, but don't have it set. so they'd have to physically edit the config file to remove the prompt.
07:00 < waxwing> and then of course after that they can change it any time.
07:03 < belcher_> yes     
07:03 < belcher_> sounds good to me
07:03 < belcher_> after we do this, im going to think carefully about how quantized order choosing should work, iv added it to my todo list now
07:06 < waxwing> i'm sketching it out, it turns into an essay on the command line ... sheesh
07:13 < waxwing> ok i wrote it out a bit at the end of the PR comments there.
07:18 < belcher_> i meant designed the quantization system so you dont need so many cli parameters
07:20 < belcher_> oh sorry you meant the random_under_limit explaination
07:23 < waxwing> heh, yeah crossed wires
07:25 < waxwing> belcher_, it suddenly occurs to me (duh) that it'd be interesting to see if there was any fee stickiness around the existing fixed defaults for -x in tumbler.
07:26 < waxwing> i guess the simple answer is no; i think fees tend to be significantly lower than those values.
07:26 < waxwing> well that also duh isn't it, because of the weighting.
07:26 < belcher_> yes     
07:27 < belcher_> its exponential so the -x values are meant to much higher than what is normally paid
07:30 < waxwing> you could try to find some correlation but it'd be so vague it'd be totally uninteresting
07:33 < belcher_> with that cli prompt you wrote, maybe this random-under-max could actually be made the default order choosing algo
07:33 < belcher_> since the cli prompt explains everything
07:34 < waxwing> yes it makes the argument for it being default more persuasive
07:34 < waxwing> but that prompt is quite messy even in that prototype form. not ideal ...
07:45 < belcher_> it can be improved, i think the general idea is very good
07:45 < belcher_> "you will should" should probably be something like "joinmarket will choose"
07:48 < waxwing> "you will choose" you mean? yeah good point, much clearer
08:12 < belcher_> whoops, yes
08:13 < belcher_> theres more to bikeshed im sure, we have to put outselves in the perspective of someone who doesnt know much about joinmarket, which iv always found hard
08:16 < waxwing> git syntax is bonkers
08:17 < waxwing> $ git rebase -Xtheirs branch-b # <- ours: branch-b, theirs: branch-a
08:17 < waxwing> $ git merge -Xtheirs branch-b  # <- ours: branch-a, theirs: branch-b
08:24 < dx25> bumping my question from a couple days ago (had znc problems subsequently)
08:24 < dx25> <davex__> are there any plans for improving the UI?  e.g. upgrade from qt4, separate gui designer code from program logic?
08:24 < dx25> <davex__> or maybe complete redesign?
08:24 < dx25> <davex__> also things like supporting makers
08:25 < belcher_> i dont have plans for those things, but its an open source project so maybe other people do
08:26 < dx25> was thinking about working on it myself, but want to see where others are going with it first (if anywhere).
08:28 < waxwing> dx25, yes all those things sound like a plausibly helpful thing. you might find it interesting to read docs/TODO.md to get a sense of what outstanding (bigger) issues exist.
08:28 < dx25> i've never successfully used the current UI myself, command line scripts are much easier at the moment.
08:28 < dx25> ok     
08:28 < waxwing> it seems to me the biggest one is to change to python3; the problem is that it'll require some big changes.
08:29 < waxwing> because we're using at least 1 or 2 (admittedly obscure) packages that don't work. and it's hard work anyway.
08:29 < waxwing> plus there's the backend bitcoin stuff, which really needs to switch to a binding that's being maintained (which probably means coincurve replacing secp256k1)
08:30 < dx25> you mean secp256k1-py package?  or the c++ project?
08:30 < waxwing> and then there's overall design issues. if you could make the Qt better then i guess that'll be mostly orthogonal, so go for it if you want :)
08:30 < waxwing> and although it's not perfectly extricated from the business logic, it's *a lot* better than it was for that, believe me!
08:31 < waxwing> i worked very hard on it, never having done anything similar in the past, it was very difficult really. but i'm not inclined to spend huge amounts of time trying to improve it, given (a) all the other things i'd like to work on and (b) all the other things (as mentioned above) that joinmarket rather badly needs.
08:31 < waxwing> dx25, yes secp256k1-py
08:32 < dx25> i had looked a little into using qt5 designer, or maybe even the newer QML stuff with qt creator.  have you tried either of those?
08:35 < waxwing> well that stuff'd be nice i guess, could clean up the UI aspects, so please do try it out if you like, but it's probably more basic things that most need improving. for example, as on TODO.md, having a binary build process that works cross-platform, plus glitches like: the wallet sync is supposed to be asynchronous, but isn't right now.
08:35 < waxwing> adding better UI while tumbler runs.
08:35 < waxwing> btw last time i tried it (a few months ago) i could do a full tumbler run on it and it worked. but .. i'm not the best tester to be fair :)
08:35 < waxwing> so if you tell me it doesn't work, i guess i have to accept that that is not unlikely.
08:35 < dx25> no i didn't mean that,
08:36 < dx25> i just already knew how to use the scripts
08:36 < waxwing> oh yes scripts is better; not much question about that.
08:36 < dx25> so i didn't do much with it.
08:37 < dx25> well i'll take a look at that todo before i work on anything.  thanks.
08:37 < waxwing> adding makers as you suggest, would clearly be a cool feature. i'd think makers are particularly likely to be in that set of people that are comfortable on the command line anyway, though.
08:37 < dx25> probably
08:38 < waxwing> dx25, you could also look at the Issues on the github marked 'Enhancement' if you'd like to contribute, we'd be grateful :)
08:38 < dx25> ok     
08:59 < belcher_> the proportion of nodes not vulnerable to the inflation bug CVE-2018-17144 has gone above 25% according to luke-jr's site
09:00 < belcher_> so umm.. yay!
09:02 < waxwing> heh     
09:15 -!- azizLIGHT [~azizLIGHT@unaffiliated/azizlight] has joined #joinmarket     
09:27 -!- felix34 [~muh@5.2.76.19] has joined #joinmarket     
09:29 < felix34> did te latest round of `apt update && apt upgrade` break joinmaket-clientserver for everyone or just me? am now getting a `jmclient.jsonrpc.JsonRpcError` when sync_wallet() is called from blockchaininterface.py
09:31 < felix34> (bitcoind is now at 0.17)
09:32 < waxwing> felix34, if you're not using master, you need to start bitcoind with --deprecatedrpc=accounts
09:32 < waxwing> that is, if you just updated bitcoin from 0.16 to 0.17
09:32 < waxwing> if you're on master, it'll work without that extra flag
09:38 < felix34> waxwing: yep, I am using a fairly old version actually. I should probably upupgrade, is there anything I need to know before doing `git pull` other than "backup yout wallet.jmdat"?
09:38 < undeath> dx25: working on the gui code seems like a good idea. Qt4 is EOL, so any effort to remove that is certainly not waisted
09:38 < undeath> *wasted
09:38 < waxwing> felix34, do you currently have a wallet.jmdat?
09:39 < waxwing> so you pulled from master some weeks ago, not using a release
09:39 < waxwing> in that case, no, nbd, just git pull should be OK
09:42 < felix34> waxwing: yes, currently have a wallet, pulled from master a few months ago
09:43 < undeath> the last release was still using wallet.json
09:44 < undeath> if you already have a wallet.jmdat you should absolutely update to the latest version on master
09:54 < felix34> Hmm, I updated to the latest master AND added `-deprecatedrpc=accounts` to the systemd unit file for bitcoind, and restarted, but now I am just getting JsonRpcErrors in a different place: blockchainInfo = self.jsonRpc.call("getblockchaininfo", [])
09:55 < undeath> don't do both
09:57 < felix34> Ok, took out deprecatedrpc and restarted bitcoind, but error persists
09:57 < undeath> can you post the full error? use https://paste.fedoraproject.org
10:02 < felix34> so the weird thing is it produced 2 slightly different erros on consecutive runs, even though all that happened was bitcoin continuing its startup process. 
10:02 < felix34> https://hashb.in/#BQKwtgbgpgdhCUACArgZygJwALo9DAXAH4D0IA9gJYxgCGGA1lAC4C0AxgDaWzO74lU7DJQAOfACSJRAT2YALcjEToYAE1G0ZYXgDpZiAO61OnFrvBrazRAEZ7iAB7OAUABUMtdlABGXhojAYOSoNhhQ3jA27CaciJy0ofAELoiIAGKUZogARKoaWjpR+jI5ADTx1FCIAEwAzABsFdSIADzBashmAHypaYh01MDwfZnZebAF2nqy5ZUw1QCsts3KgzB9aZzktGoA+qIY5ADmnmB77EoAZpTHw6NZ1TkkijokaJhkVDT0TGxcPCi/E+4GgcBI2xinAhlB8JFkCiUNV0AHZBJRmFBWJp2AxaMcoKgyGAAbwSJcYDdjshwiU5twFog
10:02 < felix34> 6osaqt4jt9ocTmcLtdbptEMdtn5OHtUNQhSwlLoKVSRmkxk8XuQ3h8MF9qHRGCwONxeMD1aDYBAIeQoTC4QjFDBkWjxZjsf58YTiaSouS+dTabMKgzqgAWAAMTUQLQJzD2PkhDHY8lo1D21ExGCuXigiZgoVoMG8Ap87AzydT3kQAF5EAAhDGXagAYXI4QAklFMMWoMAMKJ2BUFsxDA2GPKMo9csrVbgNT9tf89UDMAIjeDISYLfC5NbbeiHTi8QSieA3cwSFGzTG49Qk6203TfVUmbZ/Wy9hmMc+89HY/GYM2ruQyyooE4K4LFQJQACUuxlWJgBycMT1xT8L0pcg5gAbQAXSHRVR1eKB3gnChNV+HVDwNYkwRNZdoW4S11yRVEtyxHdnX3ElZyPEBQJg
10:02 < felix34> Tt2BveZqgfEMWihTgBU8Sh0EQAApLiIPYABRDAjgwDtCVEJR0FQnJMGUnJMJcA92JApQeN0GTwK7RTlJcUBIGNJA1RwedMGISctT+XVAT4Zz1SEERxFQKQrSUAD1E0aZigMYxTHMSxrDsBxnEcFx5JbDAjFiFhEDUCIMBkcRKBCzRUFQUR5E8dAUgAWSqgARUNUAINJ/XcTxvD8XFAmCUJEHCSJolieJEmYZIHnGfJwqKZg+L9WpGjZdpyE6HoBXWe4FRHCYwsKGZSlvRlbBqAAONl1gFVAZBzPZorMZhgBulgKlTUJS3IAqNN0Z6+Eu9gsM2sc8LVNziJnLyyMXSizRXGi10RG0GPtJinT3V12OPD9zxgS8U2vH1+PvNkLquh7mAFcAxQlMxhl0fNCyv
10:03 < felix34> bxdCJgsSfuzKRrGpVcPwkFvnckj2PBuylyh6jYVhjcEYxJHcRY1GvPR09EKxtK2xmu86gATkWQmfuutnzuQURMGAKtmBrGB6ybFW0wqdAgPgRndZZkmKnCLMMAjdgfFLN3mHoT2fD+8YAe5w1eZBzz9R88jjVNc0YeC+G7Slx0ZZRwz5fgs8v2x1W8dm/1H1DZQmb1mLSf6ACgMdq7djUN30FQVny9dwk/Y9i5A45nCVUBgjw+nSO5zwEEhch+OxcTzdEdT3cXQzsks6V3Pcb2/H/S1nXa7UevCUbgVKDAdSPagfY64bwl/3QO67eAnjgAAcnDc+98JHwZC8S5kCiB+KlQkm9gwFoDoTCQdOa91DsDQepFo4QzjtDSedEk6MVnrLBe7ol6YxXgzfO6t/Q
10:03 < undeath> don't post so long urls :/
10:03 < felix34> rGLr1LsYlL7llvsZGAckoKmCCCwRQagKj0GOKgMBPdxw8yItAgWsCx7wNFrROG08U7MXTmxeWnETKQVwQdf0QllAiTEvGSS5lKGWSUg2VSpUNJQC0jpBsekRjoOmhI7ikEVFySsg2FwQA=
10:03 < felix34> ...ok, sorry about that, let me try a different paste site, that was a bad idea
10:04 < felix34> https://0bin.net/paste/Xi2hgRvCLbQIxyAF#NKcHcEu9EH+mMnDiQv+b2GpVGiBLDEy8VmxE4jTIUAj
10:05 < undeath> that error isn't very helpful
10:05 < undeath> is bitcoind started up correctly?
10:05 < undeath> or is it possibly still initialising?
10:07 < felix34> I agree it's not very helpful. At this point it has started up yes, and has peers connected. This is the same error as before when bitcoind had been up for days
10:09 < felix34> hmm, I just re-added `deprecatedrpc` as above on a lark and it worked
10:10 < undeath> oh, I just noticed the second trace indicates you are still running the outdated version
10:11 < felix34> $ git describe --tags
10:11 < felix34> v0.3.5-61-g697d8d7
10:11 < undeath> a simple git pull doesn't do the trick unless you installed in developer mode
10:12 < felix34> oh, then what should I do?
10:12 < undeath> otherwise you have to rerun setupall.py
10:13 < undeath> python setupall.py --client-bitcoin
10:13 < undeath> python setupall.py --client-bitcoin --daemon
10:13 < undeath> actually
10:14 < undeath> there should be an --all flag
10:14 < qubenix> oh, they can be run together (--client-bitcoin and --daemon)? i always used a &&
10:14 < undeath> no, they can't
10:14 < undeath> sorry
10:14 < undeath> just noticed
10:18 < felix34> ok, so I do `git pull && python setupall.py --client-bitcoin && python setupall.py --daemon` ?
10:21 < undeath> yes     
10:23 < qubenix> could do a `git verify-commit HEAD` before setup
10:24 < felix34> appears I need also to upgrade the config file: Exception: Config file does not contain the required option 'channel' in section 'MESSAGING'.
10:25 -!- stoner19 [stoner19@gateway/vpn/privateinternetaccess/stoner19] has joined #joinmarket     
10:25 < undeath> how would that even have worked before?
10:25 < qubenix> if you rename your config to joinmarket.cfg.bak or something and then run wallet-tool.py --help it will generate a new config with the right style
10:27 -!- GitHub177 [GitHub177@gateway/service/github.com/x-cqjlcwyfnbwdnquj] has joined #joinmarket     
10:27 < GitHub177> [joinmarket-clientserver] undeath opened pull request #203: add setupall.py --all mode (master...setupall-all) https://git.io/fx7Vl
10:27 -!- GitHub177 [GitHub177@gateway/service/github.com/x-cqjlcwyfnbwdnquj] has left #joinmarket []     
10:28 < felix34> qubenix: done, tahnks
10:28 < qubenix> np     
10:49 -!- undeath [~undeath@hashcat/team/undeath] has quit [Quit: WeeChat 2.2]     
12:16 -!- reallll [~belcher@unaffiliated/belcher] has joined #joinmarket     
12:19 -!- belcher_ [~belcher@unaffiliated/belcher] has quit [Ping timeout: 244 seconds]     
12:28 -!- undeath [~undeath@hashcat/team/undeath] has joined #joinmarket     
12:52 < luke-jr> belcher: 25% is still pretty scary :x
12:53 -!- reallll is now known as belcher_     
12:53 < belcher_> luke-jr yep, shocking really
12:56 < undeath> if you don't force updates on people like chrome or windows nobody is going to care
12:58 < waxwing> i can't remember why there wasn't an "all" method; it was something to do with (a) wanting to be able to install for a plugin (so no bitcoin) (b) wanting to be able to install daemon only. but .. why not include "all"? can't remember any good reason.
12:59 < undeath> :)     
12:59 < undeath> I guess a simple oversight because the install.sh script will take care of it
13:00 < undeath> but if you later only need to update the jm source it's handy to have
13:37 < belcher_> im watching the lightning hack day video
13:38 < belcher_> joinmarket is briefly discussed during the q&a at the end of alex bosworth's talk
13:38 < waxwing> ah yeah i was mentioning that the other day in #lnd ; overall he makes some pretty sound comments about fungibility tech
13:41 < belcher_> yep     
13:43 < luke-jr> undeath: under that assumption, Bitcoin cannot work
13:43 < luke-jr> [19:56:07] <undeath> if you don't force updates on people like chrome or windows nobody is going to care
13:45 -!- stoner19 [stoner19@gateway/vpn/privateinternetaccess/stoner19] has quit [Ping timeout: 252 seconds]     
14:06 -!- stoner19 [stoner19@gateway/vpn/privateinternetaccess/stoner19] has joined #joinmarket     
14:11 -!- stoner19_ [stoner19@gateway/vpn/privateinternetaccess/stoner19] has joined #joinmarket     
14:14 -!- stoner19 [stoner19@gateway/vpn/privateinternetaccess/stoner19] has quit [Ping timeout: 252 seconds]     
14:18 -!- stoner19_ [stoner19@gateway/vpn/privateinternetaccess/stoner19] has quit [Ping timeout: 245 seconds]     
14:20 -!- stoner19 [stoner19@gateway/vpn/privateinternetaccess/stoner19] has joined #joinmarket     
14:41 -!- stoner19_ [stoner19@gateway/vpn/privateinternetaccess/stoner19] has joined #joinmarket     
14:42 -!- stoner19 [stoner19@gateway/vpn/privateinternetaccess/stoner19] has quit [Ping timeout: 264 seconds]     
14:54 < waxwing> lol on cgan i get this:
14:54 < waxwing> Tor_44> hello
14:54 < waxwing> <Tor_44> !swabsoffer 0 100000 281462636 0 1000 ~
14:54 < waxwing> <Tor_44> you have material about ethical hacking
14:54 < waxwing> perhaps i already failed at being an ethical hacker since that was a PM. sorry :)
14:55 < belcher_> every so often you get actual humans coming into that channel being a bit confused
14:55 < belcher_> it says what it is on the /topic
14:55 < waxwing> right, but that's extra special confused. is he hedging whether i'm a bot or human? :)
14:57 < belcher_> :p     
15:55 < undeath> waxwing: I can merge #203?
16:51 -!- undeath [~undeath@hashcat/team/undeath] has quit [Quit: WeeChat 2.2]     
17:09 -!- M1 [~Michail@michail.com] has quit [Ping timeout: 264 seconds]     
17:12 -!- M1 [~Michail@michail.com] has joined #joinmarket     
17:55 -!- stoner19_ is now known as stoner19     
18:22 -!- stoner19 [stoner19@gateway/vpn/privateinternetaccess/stoner19] has quit [Excess Flood]     
18:34 -!- AgoraRelay [~jmrelayfn@p5DE4ADBF.dip0.t-ipconnect.de] has quit [Ping timeout: 272 seconds]     
18:47 -!- AgoraRelay [~jmrelayfn@p5DE4ADDE.dip0.t-ipconnect.de] has joined #joinmarket     
20:05 -!- felix34 [~muh@5.2.76.19] has quit [Quit: leaving]