--- Day changed Tue Sep 03 2019
01:05 -!- deafboy [quasselcor@cicolina.org] has quit [Quit: deafboy]     
01:05 -!- deafboy [quasselcor@cicolina.org] has joined #joinmarket     
01:05 -!- deafboy [quasselcor@cicolina.org] has quit [Client Quit]     
01:05 -!- deafboy [quasselcor@cicolina.org] has joined #joinmarket     
04:10 -!- MaxSan [~four@195.206.105.227] has quit [Quit: Leaving.]     
06:12 < waxwing> kristapsk, yes thanks it's a good point that the automated tests should be mentioned, and perhaps prominently. i do still feel that what is on that page ought to be extremely useful to people who actually want to develop on joinmarket, because there is a lot (especially holistic 'integration' testing) that isn't at all covered by the automated tests.
06:14 < waxwing> a bit like regtest with Bitcoin you need to be able to see the full workflow between bots. now of course we could put a lot more in the automated testing that does that, but that hasn't been done and i think that kind of manual testing is always going to be necessary unless a *huge* amount of work is done to create tests addressing all combinations of behaviours.
06:14 < waxwing> so anyway thanks if you add stuff about run_tests, that will be very helpful.
06:15 < waxwing> i am (unfortunately this was forced on me) travelling today and tomorrow (currently at airport). i can't really address the bug that lopp found i'd appreciate it if other people tried to reproduce.
07:44 -!- davex__ [~user@45.74.60.132] has quit [Ping timeout: 246 seconds]     
07:44 -!- davex_ [~user@45.74.60.132] has quit [Ping timeout: 245 seconds]     
08:16 -!- davex__ [~user@97.119.168.55] has joined #joinmarket     
08:16 -!- davex_ [~user@45.74.60.152] has joined #joinmarket     
08:20 -!- davex__ [~user@97.119.168.55] has quit [Ping timeout: 245 seconds]     
08:21 -!- davex__ [~user@45.74.60.152] has joined #joinmarket     
08:21 -!- sosthene [~sosthene@88.191.20.124] has joined #joinmarket     
10:14 < waxwing> so pyaes doesn't have GCM .. hmm
10:23 < CgRelayBot> [cgan/AlexCato1] @arubi , do you maybe have an idea what might be the problem of JM installation in that case: https://old.reddit.com/r/joinmarket/comments/cytsln/install_problems/
10:24 < CgRelayBot> [cgan/AlexCato1] basically: user is running Ubuntu and tries to run install.sh
10:24 < CgRelayBot> [cgan/AlexCato1] error:
10:24 < CgRelayBot> [cgan/AlexCato1] /joinmarket-clientserver-master/install.sh: line 332: pushd: jmbase: No such file or directory
10:24 < CgRelayBot> [cgan/AlexCato1] ERROR: Directory '.' is not installable. Neither 'setup.py' nor 'pyproject.toml' found.
10:27 < takinbo> making some progress with refactoring the installation process. got a few questions about the current installer. 
10:28 < takinbo> first, is openssl used anywhere within joinmarket?
10:29 < takinbo> secondly, was there a specific reason for choosing the version of libsecp256k1?
10:30 < takinbo> coincurve is pinned at 0.9.0, any specific reason? i checked and there are more recent versions that have improvements that could be leveraged.
10:31 < takinbo> so far, I've gotten everything I've tested to run with the only external dependency being libsodium. 
10:32 < waxwing> takinbo, unfortunately openssl is still a dependency of twisted.
10:32 < waxwing> libsecp is coming just from coincurve yeah.
10:33 < waxwing> re: pin. i think the answer is no. currently at airport will try to check soonish.
10:34 < takinbo> ok. i suspect i didn't have to install openssl because it is installed by default on most linux distros.
10:34 < waxwing> if we pin to an old libsodium that is also something we can/should review. (it's worthy of note we only use the most basic e2e encryption mechanism).
10:35 < takinbo> yes. since we aren't doing anything fancy, is there a reason not to use the system libsodium?
10:36 < waxwing> oh; libsodium is installed by default nowadays?! i had no idea :)
10:37 < takinbo> not by default but it is available and could be installed from the package repository without the need to build from sources. 
10:37 < waxwing> oh you mean that.
10:38 < takinbo> yes     
10:38 < waxwing> hmm yes that is certainly reasonable, i do think people like the ability to compile, but they could always do that separately. so i agree it would be good to clean that out.
10:38 < waxwing> perhaps there are other opinions.
10:39 < takinbo> would also be interested in other opinions
10:39 < CgRelayBot> [cgan/AlexCato1] i like version pinning for security reasons. Repos have been taken over in the past, if we blindly just install whatever is the most current version, this might spell trouble
10:39 < CgRelayBot> [cgan/AlexCato1] about repos from source/distro: no preference on my side
10:40 < waxwing> yes pinning is good, but otoh it's also a shame if it's not constantly updated that you might miss fixes without realising it.
10:40 < CgRelayBot> [cgan/AlexCato1] yeah, thats the downside
10:41 < waxwing> alexcato i notice he has `joinmarket-clientserver-master` .. is it perhaps to do with that? otoh the guy seems to not understand basic command line stuff so it may be very difficult to diagnose.
10:41 < CgRelayBot> [cgan/AlexCato1] basically, we need an automated trigger to tell us when there's a newer version, so we can check if the pinning can be updated and if there is security fixes (strong pro to update then) or just features (which we most likely dont need)
10:44 < CgRelayBot> [cgan/AlexCato1] oh, right, it has -master in it. Even weirder that a 'git clone' didnt work then, so he also has it installed without "-master"
11:20 < kristapsk> takinbo, wasn't the whole "building locally" thing partially to just support non-Debian OSes?
11:25 < takinbo> kristapsk: i have no idea actually. considering the dependencies i've had to install (currently only libsodium) there's a good chance that other distros now have decent support for the other libraries. 
11:28 < kristapsk> the cool thing about install.sh is that you don't need to do additional manual steps, it just downloads and compiles locally
11:28 < takinbo> i'm in support of maintaining the option to install from source as much as we can; it's just that we rely on a number of packages that are not built from source and so if it makes it easier to install and use jm, i think it's worth the effort. 
11:29 < kristapsk> (probably that should be mentioned in docs somewhere tough, when I set up joinmarket-cs for the first time, I didn't ran install.sh, because was afraid it will install something system wide)
11:29 < kristapsk> I once wanted to add Gentoo support to install.sh, but decided it's not worth it, I could just compile locally for JM instead
11:31 < kristapsk> waxwing, about " the guy seems to not understand basic command line stuff" - that's why I would like to have some GUI-only install and everything at least for JoinMarketQt, not everybody uses console, even on Linux/Mac (we can learn from Wasabi how simple install / usage can be)
11:57 < arubi> AlexCato1, I'm not sure what's going on there.  Look like permissions \ out-of-space issue or some other unexpected weirdness.
11:58 < arubi> maybe the git repo he currently has on there was cloned by a different user (maybe root?) and now he can't write\remove it..  anyway, I posted on the thread.  if he's worried about a random guy telling him to run commands, you or waxwing might need to vouch for me :)
11:58 < arubi> I'm moosapor on reddit
12:02 < CgRelayBot> [cgan/AlexCato1] thanks arubi. Yeah, permission issue was my idea as well, thats why i wanted to make sure he clones as user into his home dir. Let's see what he reports back
12:21 -!- reallll [~belcher@unaffiliated/belcher] has joined #joinmarket     
12:25 -!- belcher [~belcher@unaffiliated/belcher] has quit [Ping timeout: 258 seconds]     
12:30 -!- reallll is now known as belcher     
13:25 -!- MaxSan [~four@185.156.175.35] has joined #joinmarket     
14:30 < kristapsk> waxwing, Can't get that regtest stuff working. Fixed that restart_cb line, but entering something random 32 chars long at "generate wallet" just gives me empty error message. Obviously I'm not doing stuff right, but empty error message also does not seem to be something as expected...
14:31 < kristapsk> running pytest with ygrunner.py as you did in Malta video just gives me bunch of "ConnectionRefusedError: [Errno 111] Connection refused" errors
14:32 < kristapsk> when I do ./test/run_tests.sh, everything passes, and I can't find what is different there so far, as it also runs pytest
14:46 < kristapsk> I think I will send PR anyway, as it clearly fixes that one issue, screwed up in #342 
14:48 < waxwing> kristapsk, should have some time for this tomorrow afternoon, will get back to you then. don't be too worried about the fact there's a problem in regtest with Qt, that's very far from urgent. we should focus on whether what lopp observed is reproducible on mainnet, because that's a really big deal if so.
14:50 < kristapsk> waxwing, agree, and this can be actually tested with testnet too, I guess, as it does not require yg's or anything
14:56 < kristapsk> anyways, I sent that PR with regtest fix
15:12 -!- kristapsk [~KK@gateway/tor-sasl/kristapsk] has quit [Remote host closed the connection]     
15:14 -!- kristapsk [~KK@gateway/tor-sasl/kristapsk] has joined #joinmarket     
15:21 -!- Zenton [~user@unaffiliated/vicenteh] has quit [Ping timeout: 245 seconds]     
15:46 -!- technonerd [~techno@gateway/tor-sasl/technonerd] has quit [Ping timeout: 260 seconds]     
15:52 -!- technonerd [~techno@gateway/tor-sasl/technonerd] has joined #joinmarket     
16:04 < kristapsk> I can confirm that jlopp's issue happens for me with blank testnet wallet too
17:06 -!- CgRelayBot [~CgRelayBo@p5DE4ACED.dip0.t-ipconnect.de] has quit [Ping timeout: 244 seconds]     
17:06 -!- AgoraRelay [~jmrelayfn@p5DE4ACED.dip0.t-ipconnect.de] has quit [Ping timeout: 246 seconds]     
17:17 -!- AgoraRelay [~jmrelayfn@p5DE4AAF0.dip0.t-ipconnect.de] has joined #joinmarket     
17:19 -!- CgRelayBot [~CgRelayBo@p5DE4AAF0.dip0.t-ipconnect.de] has joined #joinmarket     
22:32 < waxwing> and with tag 0.5.4?