--- Day changed Sun Aug 19 2018 01:43 -!- undeath [~undeath@unaffiliated/undeath] has joined #joinmarket 04:01 -!- mode/#joinmarket [-r] by waxwing 04:04 <@waxwing> i'm going to merge the big one now (104). obv wanted feedback from others but none is forthcoming, so there it is. also note i'm going to add a couple of minor patch-up commits after, more or less immediately, so if you want to run it from master wait a short while. 04:06 -!- GitHub44 [GitHub44@gateway/service/github.com/x-buvwejiqsagylixm] has joined #joinmarket 04:06 < GitHub44> [joinmarket-clientserver] AdamISZ closed pull request #104: [WIP] re-implement wallet (master...new-wallet) https://git.io/vFcig 04:06 -!- GitHub44 [GitHub44@gateway/service/github.com/x-buvwejiqsagylixm] has left #joinmarket [] 04:06 -!- GitHub134 [GitHub134@gateway/service/github.com/x-cbfufrnyypgmgwuj] has joined #joinmarket 04:06 < GitHub134> [joinmarket-clientserver] AdamISZ pushed 24 new commits to master: https://git.io/fAv33 04:06 < GitHub134> joinmarket-clientserver/master ca57a14 undeath: add new wallet implementation 04:06 < GitHub134> joinmarket-clientserver/master 474a77d undeath: add setup.py dependencies 04:06 < GitHub134> joinmarket-clientserver/master 995c123 undeath: replace old wallet implementation with new one 04:06 -!- GitHub134 [GitHub134@gateway/service/github.com/x-cbfufrnyypgmgwuj] has left #joinmarket [] 04:06 < undeath> waxwing: I was just about to fix the release notes :) 04:07 <@waxwing> nbd 04:07 <@waxwing> you weren't around yesterday otherwise i would have asked you 04:08 < undeath> yeah, sorry, didn't have any time yesterday 04:11 -!- GitHub99 [GitHub99@gateway/service/github.com/x-frbcolumyhiplvzx] has joined #joinmarket 04:11 < GitHub99> [joinmarket-clientserver] AdamISZ pushed 1 new commit to master: https://git.io/fAv30 04:11 < GitHub99> joinmarket-clientserver/master 455d092 AdamISZ: minor bugfixes to Qt for new wallet code 04:11 -!- GitHub99 [GitHub99@gateway/service/github.com/x-frbcolumyhiplvzx] has left #joinmarket [] 04:11 < arubi> nice undeath :) 04:11 * arubi rebases stuff.. 04:11 < undeath> big thanks to waxwing for all the testing! 04:12 <@waxwing> i'm going to add one more: delete the jmtainter.py script; not even sure whether it works or not, but have no intention of trying to maintain it right now, so def should be done i think 04:12 < undeath> I'll prepare a pr to address the issues you mentioned in your last comment 04:13 <@waxwing> np, i'm actually pretty happy with what undeath 's done in terms of it being a cleaner design, more flexible, better security model for the wallet etc. ... it'll not impact users much at all, but i guess that's why it's all the more important to get the communication/doc as clear as possible, as they could be easily confused about what the change represents. 04:14 < undeath> thanks :) 04:14 < undeath> yep 04:15 -!- GitHub39 [GitHub39@gateway/service/github.com/x-wqlldcdnfjttasfe] has joined #joinmarket 04:15 < GitHub39> [joinmarket-clientserver] AdamISZ pushed 1 new commit to master: https://git.io/fAv3r 04:15 < GitHub39> joinmarket-clientserver/master 55c51a9 AdamISZ: Remove jmtainter script, not used or maintained currently 04:15 -!- GitHub39 [GitHub39@gateway/service/github.com/x-wqlldcdnfjttasfe] has left #joinmarket [] 04:22 <@waxwing> undeath, btw i remember saying i was finding it to have sped up wallet sync on a big wallet; now i can't really say that; over a long running period i'd say it's maybe about the same or perhaps a bit slower. 04:23 < undeath> ok, that's more what I feared 04:23 <@waxwing> a related point that i need to remember: the current wallet load in Qt goes into "gui freeze", i.e. is blocking during wallet sync. need to fix that. 04:24 < undeath> did that not happen with the old wallet? 04:24 <@waxwing> i think it's subtly changed from the way it worked before; i think i'd fixed that somehow, earlier. anyway that's one of a number of minor tweaks the GUI needs, not sure if i'll get around to it. 04:24 <@waxwing> yeah i can't exactly remember tbh. i'll need to investigate. 04:24 <@waxwing> prob it was like that before. 04:25 < undeath> I think there is a bug when a maker has a sw wallet and imported legacy keys. I have not seen any successful cj using one, but two or three failed ones 04:25 <@waxwing> so if anyone here wants to run from master, you can now. just run the convert_old_wallet script. 04:25 < undeath> it's probably something with the validation done in the taker 04:25 <@waxwing> undeath, ok. if you need help fixing it let me know. 04:26 < undeath> ok, thanks. I'll try to do some regtest testing and report back 04:44 -!- Lymia7 [~Lymia@node-t8d.pool-180-180.dynamic.totbb.net] has joined #joinmarket 04:44 < Lymia7> LRH OFFICIAL: We are not spamming you | https://www.youtube.com/watch?v=_utMUBnl3nk 04:46 -!- missnomer23 [~missnomer@197.58.137.166] has joined #joinmarket 04:46 < missnomer23> LRH OFFICIAL: We are not spamming you | https://www.youtube.com/watch?v=_utMUBnl3nk 04:47 -!- Lymia7 [~Lymia@node-t8d.pool-180-180.dynamic.totbb.net] has quit [Remote host closed the connection] 04:48 -!- GitHub82 [GitHub82@gateway/service/github.com/x-hrhefaarberrebuv] has joined #joinmarket 04:48 < GitHub82> [joinmarket-clientserver] undeath opened pull request #181: new wallet follow-up (master...wallet-usability-fix) https://git.io/fAvs6 04:48 -!- GitHub82 [GitHub82@gateway/service/github.com/x-hrhefaarberrebuv] has left #joinmarket [] 04:53 -!- missnomer23 [~missnomer@197.58.137.166] has quit [Remote host closed the connection] 04:56 -!- nickenchuggets [~nickenchu@49.134.201.77.rev.sfr.net] has joined #joinmarket 04:59 -!- nickenchuggets [~nickenchu@49.134.201.77.rev.sfr.net] has quit [Remote host closed the connection] 05:00 <@waxwing> lol that's a new one "we are not spamming you" 05:00 <@waxwing> thanks undeath 05:06 < undeath> whoops, forgot a .format(), force-pushed 05:24 -!- change [~change@KD027082173212.ppp-bb.dion.ne.jp] has joined #joinmarket 05:24 < change> LRH OFFICIAL: We are not spamming you | https://www.youtube.com/watch?v=_utMUBnl3nk 05:27 -!- change [~change@KD027082173212.ppp-bb.dion.ne.jp] has quit [Killed (Sigyn (Spam is off topic on freenode.))] 05:34 < undeath> ok, after some testing I cannot find a problem with mixed maker wallets. There may be a problem with the checks in the old taker code. 05:34 -!- mode/#joinmarket [+r] by waxwing 05:34 <@waxwing> if anyone has a better setting for anti-spam i'm willing to try it 05:41 <@waxwing> arubi, so seems like #180 is mergeable then? 05:47 < undeath> oh, now that -qt uses open_test_wallet_maybe the change in #181 is bad 05:47 < undeath> I guess I'll just revert the changes to open_test_wallet_maybe 05:47 < undeath> I added those so the user does not see the exception stack trace when something goes wrong 05:48 <@waxwing> undeath, ah hang on, let me take a look, haven't read it yet 05:49 <@waxwing> yes, this has been a pain point quite a few times, going back to the start of the project i guess: triggering different behaviour in testnet or regtest mode using kind of hacky things like arguments and whatnot. 05:50 < arubi> waxwing, yep, #180 is ready 05:50 <@waxwing> i'd rather not have to change it again right now. 05:50 <@waxwing> arubi, gotcha thanks will merge now 05:51 < undeath> I do think it's not a good idea to call sys.exit anyway 05:51 < undeath> so yeah, reverting that change is the best thing overall 05:52 <@waxwing> undeath, ack on the language of the error message 05:54 < undeath> reverted + force-pushed 05:57 < undeath> regarding moving to ~/.joinmarket, I was thinking the best way might be to have different joinmarket profiles in subfolders 05:58 < undeath> ie ~/.joinmarket/default/{wallet.jmdat,joinmarket.cfg,logs/,cmtdata/} 05:58 <@waxwing> oh hmm, never thought of that. that could help with e.g. testing. 05:59 < undeath> because people will likely want to have different joinmarket.cfg's for different wallets 05:59 <@waxwing> the way i had it for coinswapcs was ~/.appname/{appname.cfg, wallets/, logs/, cmtdata/} 05:59 <@waxwing> yeah i can believe that 05:59 <@waxwing> it's a bit more complicated ofc but .. i guess. 05:59 < undeath> also if you run multiple jm instances and only have one file for commitments the two instances may have race conditions 06:00 <@waxwing> yes. 06:01 <@waxwing> it caused slight problems in the past when simulating big trading pits. i can't remember the exact problems i saw, but they were of that flavour. 06:03 <@waxwing> undeath, can you squash those 2 commits 06:03 < undeath> ok 06:05 < undeath> done 06:05 <@waxwing> thx 06:06 <@waxwing> arubi, re yesterday's convo, i just noticed the irc test file comments out *both* the `with pytest.raises` lines; the first one is a plain old assertionerror. something must be so broken there ... oh well, another time. 06:07 < arubi> lemme see 06:08 < arubi> waxwing, you mean the first one that we taled about in junk_fill() and the second one at junk_longmsgs() ? 06:08 < arubi> s/taled/talked/ 06:10 <@waxwing> yeah junk long msgs has a commented out assertionerror check. that should be triggered afaik. 06:12 < arubi> I see.. well in the mean time the change to 35 in task.deferLater() fixed travis tests at least.. I was ready to do a PR for run_tests.sh and this change to enable osx testing 06:18 <@waxwing> it might be better if someone just finds a way to use twisted.trial properly :) probably need to use a deferred to trigger the end of the test (cpt obvious here; but it was like 1.5 years ago when i was trying to figure it out, and somehow i didn't, at the time) 06:29 < arubi> agreed, it would be better without the hard coded numbers 07:32 < arubi> so yea I'm able to reproduce what I thought were strictly macos irc errors on my machine just by changing the magic number. the errors change between setting eg '5' or '10' 07:33 <@waxwing> you mean the errors where the tests fail early with 'reactor unclean' or similar? sure. as you say, it's crappy hardcoded timeouts. 07:34 < arubi> right, seems to fail at different points based on the value in there 07:34 < undeath> the reactor unclean tests can be fixed in the same way as test_coinjoin 07:35 < undeath> you just need to figure out which test leaves the unclean state 07:36 < arubi> now that I have it on my local machine and not only on travis, this might be simpler (but still, not sure I could find it) :) 07:36 <@waxwing> yeah but the tests are cutting out early undeath ; the error is not just "there's something dangling" but "the test didn't actually complete before the reactor was shut down" 07:37 < undeath> oh 07:37 <@waxwing> this is a travis problem btw. arubi encountered it yesterday for osx travis 07:38 < undeath> I see, travis only… that's weird 07:38 < undeath> or osx only maybe 07:38 < arubi> it's not a travis only problem anymore now that I'm changing the number in there 07:39 <@waxwing> undeath, not so weird, i think the infrastructure is slower. but hmm yeah this is not really resource intensive. 07:39 < arubi> seems to be that "22" is just good enough for linux, but not for osx (on travis at least) 07:39 <@waxwing> might be to do with start up of miniircd or something 07:39 < undeath> are the tests even using the ircd? 07:39 <@waxwing> that one is, yes 07:39 <@waxwing> it's an irc test :) 07:39 < undeath> ah :) 07:40 < arubi> from the error messages, it seems that miniircd is running 07:40 <@waxwing> i was just trying to find some possible reason for it being slower. it's clearly running. 07:40 < arubi> ah, alright 08:03 <@waxwing> but prob irrelevant, eh, miniircd is started up at the start of the whole test suite, unless there's some other "starting" process i'm forgetting about. 08:06 < arubi> subprocess is imported in test_irc_messaging.py but not used (along with signal, os, and some others). maybe it used to be launched from that test but no longer 08:10 <@waxwing> oh stale imports. sorry i've never got round to cleaning that up. if someone wants to remove those that'd be great. 08:11 < arubi> I bet there's a python script that does that :P 08:13 <@waxwing> yeah prob in IDEs of various descriptions 09:12 -!- undeath [~undeath@unaffiliated/undeath] has quit [Quit: WeeChat 2.1] 10:17 < arubi> in miniircd, should I be seeing the pub\priv messages the peers send to each other if I have --debug on? 10:17 < arubi> (outside the context of the tests) 10:19 < arubi> ah yes I do see them. 10:19 < arubi> [127.0.0.1.39148:39148] -> 'PRIVMSG #joinmarket-pit-test :!orderbook' 10:19 < arubi> [127.0.0.1.39150:39150] <- ':irc_publisher!irc_publisher@127.0.0.1.39148 PRIVMSG #joinmarket-pit-test :!orderbook' 10:19 < arubi> [127.0.0.1.39148:39148] -> 'PRIVMSG #joinmarket-pit-test :!orderbook' 10:19 < arubi> [127.0.0.1.39148:39148] -> '' 10:19 < arubi> Disconnected connection from 127.0.0.1.39148:39148 (EOT). 10:19 < arubi> [127.0.0.1.39150:39150] <- ':irc_publisher!irc_publisher@127.0.0.1.39148 PRIVMSG #joinmarket-pit-test :!orderbook 10:19 < arubi> :irc_publisher!irc_publisher@127.0.0.1.39148 QUIT :EOT' 10:20 < arubi> then the tests fail and "junk and crapjunk and crapjunk..." is seen in pytest logs. interestingly, it's not on irc logs. that orderbook! is the last message 10:38 < arubi> so my kinda guess is, the next junk callback, junk_longmessages isn't happening because the server is already killed by that point 10:43 < arubi> seems like I can more messages to appear in irc logs and pytest results if I replace eg junk_pubmsgs with junk_fill in 'd = task.deferLater(reactor, 3.0, junk_pubmsgs, mc)' 10:45 < arubi> seems that orderbook junk, and both long and short "junk and crap" messages are posted to pytest debug now (so 3 out of 4), and the last message on irc log is the junk_fill one 10:46 < arubi> ah but it's not the last message that junk_fill produces: "PRIVMSG #joinmarket-pit-test :!reloffer stuff" is the last in irc logs 10:46 < arubi> (just thinking outloud in case any of this matters) 10:53 < arubi> might be missing things by tailing the logs.. I don't know if I'm reading this correctly 11:22 -!- belcher_ [~belcher@unaffiliated/belcher] has joined #joinmarket 11:59 -!- rdymac [uid31665@gateway/web/irccloud.com/x-stsyxatyjpgexdut] has joined #joinmarket 15:48 -!- belcher_ [~belcher@unaffiliated/belcher] has quit [Quit: Leaving] 16:14 -!- grubles [~grubles@unaffiliated/grubles] has quit [Ping timeout: 240 seconds] 16:15 -!- grubles [~grubles@unaffiliated/grubles] has joined #joinmarket 16:28 -!- grubles [~grubles@unaffiliated/grubles] has quit [Quit: Leaving] 22:40 -!- rdymac [uid31665@gateway/web/irccloud.com/x-stsyxatyjpgexdut] has quit [Quit: Connection closed for inactivity] 23:28 -!- instagibbs [~instagibb@pool-100-15-122-172.washdc.fios.verizon.net] has quit [Ping timeout: 240 seconds] 23:28 -!- instagibbs [~instagibb@pool-100-15-122-172.washdc.fios.verizon.net] has joined #joinmarket