--- Day changed Sun Sep 09 2018 02:03 -!- undeath [~undeath@unaffiliated/undeath] has joined #joinmarket 03:44 -!- undeath [~undeath@unaffiliated/undeath] has quit [Quit: WeeChat 2.2] 06:38 -!- undeath [~undeath@hashcat/team/undeath] has joined #joinmarket 07:18 < belcher> arubi here: https://pastebin.com/pVMHcty9 07:18 < belcher> it says U, the LD_LIBRARY_PATH seems to be empty 07:23 -!- belcher_ [~belcher@unaffiliated/belcher] has joined #joinmarket 08:32 -!- lnostdal [~lnostdal@77.70.119.51] has quit [Quit: https://www.Quanto.ga/] 08:34 -!- lnostdal [~lnostdal@77.70.119.51] has joined #joinmarket 08:34 < undeath> is the spam attack still going on or can we -r? 08:43 -!- grubles [~grubles@gateway/tor-sasl/grubles] has quit [Ping timeout: 256 seconds] 09:19 -!- undeath [~undeath@hashcat/team/undeath] has quit [Quit: WeeChat 2.2] 10:09 -!- mode/#joinmarket [+o waxwing] by ChanServ 10:09 -!- mode/#joinmarket [-r] by waxwing 10:09 -!- mode/#joinmarket [-o waxwing] by ChanServ 10:31 -!- grubles [~grubles@gateway/tor-sasl/grubles] has joined #joinmarket 10:34 < waxwing> belcher, obv i'm just guessing here, but the only thing i can think that fits the symptoms is: you have an existing, older version of the lib installed to system; during the install.sh process, attempting to download the new version fails/is not used due to detection of the old version on the system, and the old version is installed to the virtualenv. 10:40 < belcher_> ill try uninstalling it then 10:50 < waxwing> so #189 is trivial, i'll wait a day or two before merging, just in case anyone wants to take a look (but really it's nothing) 10:50 < waxwing> i've just had a read through #183, and i think there's nothing controversial (it's almost entirely fixing unused imports, and adding a flake test for that for the future on travis). 10:51 < waxwing> but again i'll wait a short while just in case. that one's been around for a week + anyway 10:51 < waxwing> well nearer 3 weeks 10:52 < waxwing> belcher_, 's #186 (accounts deprecation) is a far more important one, and def needs people to test it, so if you can please do 10:57 < waxwing> same comment for #190 as for #189, it's very clear so just waiting in case someone has a Q 11:45 < belcher_> uninstalling secp256k1 seemed to fix it (i.e. deleting the files because it seems it wasnt installed with pip) 11:46 < belcher_> also this comment points out that the calls to `listtransactions` also need to be changed https://github.com/JoinMarket-Org/joinmarket-clientserver/issues/185#issuecomment-419730061 11:46 < belcher_> i think 11:47 < belcher_> yes just confirmed it, the earlier accounts-filtering param MUST be set to * now 11:48 < belcher_> this could be pretty problematic actually, it looks like _yield_transactions() relies on the account filtering.. hmm 11:51 < waxwing> belcher_, "yield_transactions()"? 11:51 < belcher_> https://github.com/JoinMarket-Org/joinmarket-clientserver/blob/36c3fd9c57b834d2b4e687d9ed9e83d9381896e6/jmclient/jmclient/blockchaininterface.py#L592 11:53 < waxwing> oh just found it, it was part of undeath's wallet changes (sync changes) 11:56 < belcher_> looks like the results of yield_transactions() later uses wallet.is_known_addr(), which means any transactions not belonging to the joinmarket wallet just get ignored 11:56 < belcher_> ill keep reading, im mostly thinking out loud here 12:03 < waxwing> so the call to add_tx_notify in client_protocol.py specifies the wallet_name which gets passed into the listtransactions call 12:04 < waxwing> whereas the one in taker.py doesn't. i presume the former one can also be removed, meaning that in both cases the call to listtransactions uses '*' 12:04 < waxwing> does this interact with the multiwallet feature? 12:06 < belcher_> it doesnt interact at all with multiwallet 12:08 < belcher_> confusingly, "wallet name" is used for both the account/label and the rpcwallet when using the multiwallet feature 12:10 < waxwing> so, hmm, if the user's Core wallet is heavily used, we could potentially encounter problems with syncing if we use listtransactions, perhaps? i wonder why labels can't be passed to listtransactions? 12:10 < waxwing> or perhaps there's a new api for that 12:15 < belcher_> i think it would only slow it down 12:15 < belcher_> because the code after the call to listtransactions filters out the transactions that dont belong to joinmarket's wallet 12:20 < arubi> so the libsecp issue was indeed having different version (specifically, one without the pubkey recovery module) installed 12:21 < arubi> belcher_, I bet you installed it using `apt-get install libsecp256k1-{0,dev}` 12:21 < belcher_> yeah could be 12:21 < belcher_> it was a long time ago :z 12:21 < arubi> maybe try running `apt-cache policy libsecp256k1-{0,dev}` , see if any say "Installed: " 12:21 < belcher_> nope 12:22 < arubi> huh, so maybe installed from source then :) 12:22 < belcher_> its possible i checked out the git repos and installed it from there, but i checked that earlier and it didnt seem to be it 12:22 < belcher_> yes 12:22 < belcher_> maybe i installed it then deleted the git directory 12:22 < arubi> yea 12:23 < belcher_> i remember having trouble with secp256k1 back then and tried a few ways of installing it, so who knows what actually happened in the end 12:24 < arubi> that's why I love virtualenv. everything is confined nicely 12:24 < belcher_> but in this case it looked like virtualenv found the secp256k1 library from outside? 12:24 < belcher_> or is that normal 12:25 < belcher_> since otherwise you'd have to install everything again inside the venv 12:25 < arubi> yes that's normal. if it's installed system wide it probably takes precedence 12:27 < arubi> really it depends on how paths are set up. for this libsecp case I guess nothing in the secp256k1-py install process is set to ignore the system package. actually I'll look into that. maybe we can pass some flag to the setup.py installer so it avoids this in the future 12:35 < waxwing> yes it seems pretty clear that something is wrong if the installation doesn't ignore the system-wide install of the library, or at least can do so, by choice. i'd have thought it should ignore by default but i guess that's arguable. 12:36 < waxwing> belcher_, is your talk on the 19th? 12:36 < belcher_> yes 12:37 < waxwing> thx 12:45 < belcher_> for any lurkers watching, thats this https://www.meetup.com/London-Bitcoin-Devs/events/254172468/ 16:15 -!- belcher_ [~belcher@unaffiliated/belcher] has quit [Quit: Leaving] 17:49 -!- AgoraRelay [~jmrelayfn@p5DE4AC24.dip0.t-ipconnect.de] has joined #joinmarket 19:38 -!- grubles [~grubles@gateway/tor-sasl/grubles] has quit [Ping timeout: 256 seconds]