12018-10-02T00:49:03  *** ghost43 has quit IRC
  22018-10-02T00:53:33  *** ghost43 has joined #bitcoin-core-dev
  32018-10-02T01:27:26  *** Chris_Stewart_5 has quit IRC
  42018-10-02T02:15:02  *** rh0nj has quit IRC
  52018-10-02T02:16:07  *** rh0nj has joined #bitcoin-core-dev
  62018-10-02T02:16:28  *** Krellan has quit IRC
  72018-10-02T02:17:39  *** Krellan has joined #bitcoin-core-dev
  82018-10-02T02:28:15  *** Victor_sueca has quit IRC
  92018-10-02T02:29:25  *** Victor_sueca has joined #bitcoin-core-dev
 102018-10-02T02:37:25  *** Murch has quit IRC
 112018-10-02T02:50:39  *** rex4539 has quit IRC
 122018-10-02T04:04:20  *** ghost43 has quit IRC
 132018-10-02T04:04:35  *** ghost43 has joined #bitcoin-core-dev
 142018-10-02T04:14:44  *** jarthur has quit IRC
 152018-10-02T04:27:11  *** murrayn has quit IRC
 162018-10-02T04:40:41  *** gribble has quit IRC
 172018-10-02T04:42:28  *** dbnoch has quit IRC
 182018-10-02T04:47:37  *** intcat has quit IRC
 192018-10-02T04:50:04  *** intcat has joined #bitcoin-core-dev
 202018-10-02T04:53:13  *** gribble has joined #bitcoin-core-dev
 212018-10-02T05:18:38  *** Emcy has quit IRC
 222018-10-02T05:19:15  *** Emcy has joined #bitcoin-core-dev
 232018-10-02T06:13:32  <gwillen> achow101: is it intended to be the case that PSBTs always end with 0x00 (separator after the last map), or that it is always omitted, or that it is optional?
 242018-10-02T06:13:46  <gwillen> the BIP reads like it should be mandatory, but I noticed Core is inconsistent about it
 252018-10-02T06:14:15  <gwillen> (or sipa or whoever else has an opinion :-) )
 262018-10-02T06:51:05  *** hebasto has joined #bitcoin-core-dev
 272018-10-02T07:01:18  *** ken2812221 has quit IRC
 282018-10-02T07:04:30  *** reardencode has quit IRC
 292018-10-02T07:26:26  <jonasschnelli> cfields. macOS 0.17.0 detached signature is available https://github.com/bitcoin-core/bitcoin-detached-sigs/pull/17
 302018-10-02T07:35:06  *** murrayn has joined #bitcoin-core-dev
 312018-10-02T08:01:42  *** setpill has joined #bitcoin-core-dev
 322018-10-02T08:01:56  *** StopAndDecrypt has quit IRC
 332018-10-02T08:04:12  *** StopAndDecrypt has joined #bitcoin-core-dev
 342018-10-02T08:06:36  *** timothy has joined #bitcoin-core-dev
 352018-10-02T08:13:14  <wumpus> gwillen: good question; some people are travelling, if you don't get a response here, please create an issue
 362018-10-02T08:14:03  <wumpus> jonasschnelli yay
 372018-10-02T08:18:27  *** SopaXorzTaker has joined #bitcoin-core-dev
 382018-10-02T08:20:51  *** SopaXorzTaker has quit IRC
 392018-10-02T08:27:31  *** Krellan has quit IRC
 402018-10-02T08:32:36  <wumpus> and yes, the tests are sort of heavy, if you find that many functional tests fail on your system (esp. related to timeouts) it's a good suggestion to run with parallelism of one
 412018-10-02T08:34:47  <sipa> gwillen: no strong opinion, but agree it should be clear and consistent
 422018-10-02T08:36:26  *** promag has joined #bitcoin-core-dev
 432018-10-02T09:03:10  *** phwalkr has joined #bitcoin-core-dev
 442018-10-02T09:05:57  *** SopaXorzTaker has joined #bitcoin-core-dev
 452018-10-02T09:06:22  <jonasschnelli> promag: sorry,.. haven't followed the discussion... what needs update there: https://github.com/bitcoin/bitcoin/pull/12653#issuecomment-426202288?
 462018-10-02T09:07:06  <promag> IIUC block index is not affected by -blocksdir
 472018-10-02T09:09:02  <promag> jonasschnelli: gArgs.IsArgSet("-blocksdir") ? GetDataDir() / "blocks" / "index" : GetBlocksDir() / "index"
 482018-10-02T09:11:09  <jonasschnelli> promag: I see. Done. thanks.-
 492018-10-02T09:11:23  <promag> np
 502018-10-02T09:11:26  <promag> ty
 512018-10-02T09:51:40  *** hebasto_ has joined #bitcoin-core-dev
 522018-10-02T09:54:41  *** Victor_sueca is now known as Victorsueca
 532018-10-02T10:55:14  *** phwalkr has quit IRC
 542018-10-02T10:55:49  *** Chris_Stewart_5 has joined #bitcoin-core-dev
 552018-10-02T11:05:00  *** setpill has quit IRC
 562018-10-02T11:07:28  *** setpill has joined #bitcoin-core-dev
 572018-10-02T11:45:14  *** Chris_Stewart_5 has quit IRC
 582018-10-02T11:56:15  *** promag has quit IRC
 592018-10-02T12:11:55  *** shesek has quit IRC
 602018-10-02T12:31:59  *** panako has quit IRC
 612018-10-02T13:08:01  *** SopaXorzTaker has quit IRC
 622018-10-02T13:11:34  *** Guyver2 has joined #bitcoin-core-dev
 632018-10-02T13:31:10  <andytoshi> gwillen: it should end in a 0x00. that was also my read of the BIP, and it's also the only way to unambiguously parse a PSBT from a stream that might have other data coming down the line
 642018-10-02T13:32:22  *** Chris_Stewart_5 has joined #bitcoin-core-dev
 652018-10-02T13:35:31  *** promag has joined #bitcoin-core-dev
 662018-10-02T13:44:30  *** ken2812221 has joined #bitcoin-core-dev
 672018-10-02T13:52:12  <achow101> gwillen: it should end in 0x00
 682018-10-02T13:52:29  <achow101> how is core inconsistent about it?
 692018-10-02T13:52:45  <achow101> I believe there's even a test vector for this. a parser should reject a psbt if it doesn't end in 0x00
 702018-10-02T14:19:35  *** belcher_ has joined #bitcoin-core-dev
 712018-10-02T14:20:56  *** michaelsdunn1 has joined #bitcoin-core-dev
 722018-10-02T14:25:21  *** promag has quit IRC
 732018-10-02T14:54:14  *** Lauda has quit IRC
 742018-10-02T14:54:53  *** Lauda has joined #bitcoin-core-dev
 752018-10-02T14:58:59  *** profmac has quit IRC
 762018-10-02T15:12:30  *** profmac has joined #bitcoin-core-dev
 772018-10-02T15:43:21  <cfields> jonasschnelli: thanks!
 782018-10-02T15:43:30  <cfields> gitian builders: detached sigs for v0.17.0 are pushed
 792018-10-02T16:00:16  *** Krellan has joined #bitcoin-core-dev
 802018-10-02T16:04:30  *** Krellan has quit IRC
 812018-10-02T16:11:43  *** owowo has quit IRC
 822018-10-02T16:11:53  *** shinohai has joined #bitcoin-core-dev
 832018-10-02T16:19:15  *** owowo has joined #bitcoin-core-dev
 842018-10-02T16:26:25  *** promag has joined #bitcoin-core-dev
 852018-10-02T17:03:29  *** Murch has joined #bitcoin-core-dev
 862018-10-02T17:13:01  *** promag has quit IRC
 872018-10-02T17:32:20  *** W4RL0RD__ has joined #bitcoin-core-dev
 882018-10-02T17:35:55  *** Lauda has quit IRC
 892018-10-02T17:36:25  *** Lauda has joined #bitcoin-core-dev
 902018-10-02T17:43:47  *** Krellan has joined #bitcoin-core-dev
 912018-10-02T17:44:34  *** ken2812221_ has joined #bitcoin-core-dev
 922018-10-02T17:46:21  *** ken2812221 has quit IRC
 932018-10-02T17:46:27  *** ken2812221_ is now known as ken2812221
 942018-10-02T17:51:56  *** Krellan has quit IRC
 952018-10-02T17:55:03  *** CubicEarth has quit IRC
 962018-10-02T17:57:21  *** CubicEarth has joined #bitcoin-core-dev
 972018-10-02T18:18:36  <phantomcircuit> wumpus, lots of the tests seem to be racey
 982018-10-02T18:18:43  <phantomcircuit> the zmq ones especially
 992018-10-02T18:20:08  *** Zenton has quit IRC
1002018-10-02T18:20:17  *** Zenton has joined #bitcoin-core-dev
1012018-10-02T18:22:25  <wumpus> phantomcircuit: yes i's true
1022018-10-02T18:23:01  <wumpus> although I think that has improved
1032018-10-02T18:23:45  <wumpus> there's one test that simply fails on very slow systems but I don't remember which one right now, but the others pass even under very crappy circumstances
1042018-10-02T18:24:00  <wumpus> (counting only the non-extended tests)
1052018-10-02T18:33:54  <phantomcircuit> wumpus, running "test_runner.py" with small changes to the timing of networking logic makes them fail randomly
1062018-10-02T18:35:35  <gwillen> andytoshi: achow101: ok, let me double-check and then try to create a minimal example of when core creates one that does not end in 0x00
1072018-10-02T18:38:06  <gwillen> crap, okay, what happened is that I dropped a character from the end of a PSBT string while copy-pasting
1082018-10-02T18:38:15  <gwillen> and decodepsbt still accepted it
1092018-10-02T18:39:30  <gwillen> as does finalizepsbt
1102018-10-02T18:51:41  *** W4RL0RD__ has quit IRC
1112018-10-02T18:52:04  *** W4RL0RD__ has joined #bitcoin-core-dev
1122018-10-02T18:55:46  *** mthiel_ has joined #bitcoin-core-dev
1132018-10-02T18:57:33  *** mthiel has quit IRC
1142018-10-02T19:00:55  *** W4RL0RD__ has quit IRC
1152018-10-02T19:02:40  *** ExtraCrispy has joined #bitcoin-core-dev
1162018-10-02T19:40:16  *** dqx has joined #bitcoin-core-dev
1172018-10-02T19:44:21  <andytoshi> does core have a fuzz-testing harness? might be worthwhile to try round-tripping these
1182018-10-02T19:47:19  <wumpus> andytoshi: yes, test_bitcoin_fuzzy.cpp
1192018-10-02T19:48:16  <wumpus> phantomcircuit: ok... any idea which specific messages cause the problem? what kind of failures?
1202018-10-02T19:48:49  <phantomcircuit> wumpus, the zmq ones especially fail
1212018-10-02T19:49:03  <phantomcircuit> the getzmqnotifications fails
1222018-10-02T19:49:26  <phantomcircuit> which im sure is just a race
1232018-10-02T19:49:48  <phantomcircuit> it's hard to diagnose cause they tend to work on my laptop but not on travis
1242018-10-02T19:49:53  <phantomcircuit> maybe i need a slower test system
1252018-10-02T20:00:37  *** brianhoffman has quit IRC
1262018-10-02T20:06:05  <wumpus> if it's just zmq I'm not *too* worried
1272018-10-02T20:07:06  <phantomcircuit> and apparently, feature_notifications
1282018-10-02T20:07:36  <phantomcircuit> with too much parallelism everything explodes cause we run out of local ports to bind to but that's not really something to easily fix
1292018-10-02T20:08:29  <achow101> gwillen: base64 has some padding characters at the end of the string if it isn't some specifc length (a multiple of some number I can't remember). Core's implementation of base64 drops those padding characters and then decodes. you're probably dropping a padding char which is thus being ignored
1302018-10-02T20:08:58  <achow101> the padding is the = characters you see at the end of base64 strings.
1312018-10-02T20:11:32  <gwillen> achow101: the character in question is an A, which is 0, so while it's not padding I can imagine there could be ambiguity
1322018-10-02T20:12:17  <achow101> hmm. can you give me a psbt that should be wrong but decodes fine?
1332018-10-02T20:12:21  <gwillen> if we're using padding consistently, we should possibly refuse to decode truncated base64
1342018-10-02T20:12:28  <gwillen> i.e. base64 that's not a multiple of four bytes
1352018-10-02T20:12:37  <gwillen> achow101: cHNidP8BAHMCAAAAAbiWoY6pOQepFsEGhUPXaulX9rvye2NH+NrdlAHg+WgpAQAAAAD/////AkBLTAAAAAAAF6kUqWwXCcLM5BN2zoNqMNT5qMlIi7+HQEtMAAAAAAAXqRSVF/in2XNxAlN1OSxkyp0z+Wtg2YcAAAAAAAEBIBNssgAAAAAAF6kUamsvautR8hRlMRY6OKNTx03DK96HAQcXFgAUo8u1LWpHprjt/uENAwBpGZD0UH0BCGsCRzBEAiAONfH3DYiw67ZbylrsxCF/XXpVwyWBRgofyRbPslzvwgIgIKCsWw5sHSIPh1icNvcVLZLHWj6NA7Dk+4Os2pOnMbQBIQPGStfYHPtyhpV7zIWtn0Q4GXv5gK1zy/tnJ+cBXu4iiwABABYAFMwmJQEz+HDpBEEabxJ5PogPsqZRAAEAFg
1362018-10-02T20:12:46  <gwillen> er, sorry, cHNidP8BAHMCAAAAAbiWoY6pOQepFsEGhUPXaulX9rvye2NH+NrdlAHg+WgpAQAAAAD/////AkBLTAAAAAAAF6kUqWwXCcLM5BN2zoNqMNT5qMlIi7+HQEtMAAAAAAAXqRSVF/in2XNxAlN1OSxkyp0z+Wtg2YcAAAAAAAEBIBNssgAAAAAAF6kUamsvautR8hRlMRY6OKNTx03DK96HAQcXFgAUo8u1LWpHprjt/uENAwBpGZD0UH0BCGsCRzBEAiAONfH3DYiw67ZbylrsxCF/XXpVwyWBRgofyRbPslzvwgIgIKCsWw5sHSIPh1icNvcVLZLHWj6NA7Dk+4Os2pOnMbQBIQPGStfYHPtyhpV7zIWtn0Q4GXv5gK1zy/tnJ+cBXu4iiwABABYAFMwmJQEz+HDpBEEabxJ5PogPsqZRAAEAF
1372018-10-02T20:13:00  <gwillen> hm, that may be too long for IRC
1382018-10-02T20:13:30  <achow101> probably
1392018-10-02T20:13:34  <gwillen> achow101: https://0bin.net/paste/e70Ze5n2HJSSHuF1#tPS6iIwRHdEOmHx85Yo3Ueaos6ERjdrHbLUFW8Z63Dg
1402018-10-02T20:13:38  <gwillen> decodes with or without the final A
1412018-10-02T20:13:56  <gwillen> without, it's technically invalid if you want a properly-padded base64 string
1422018-10-02T20:14:07  <gwillen> but if I add an = for padding it still decodes even though it should be missing the 0 at the end
1432018-10-02T20:14:19  <achow101> ok, i'll take a look
1442018-10-02T20:16:46  <achow101> yeah, i see why that happens. can I use this psbt as a test vector?
1452018-10-02T20:18:01  <wumpus> phantomcircuit: supporting UNIX sockets would be nice, I had PRs for that once
1462018-10-02T20:19:27  <wumpus> it made the tests do RPC over UNIX sockets
1472018-10-02T20:20:16  <wumpus> idea was to do the same for the P2P connections between the nodes, which would make the tests completely self-contained
1482018-10-02T20:20:36  <wumpus> but... just one of the many of my projects I lost track of
1492018-10-02T20:21:31  <wumpus> wouldn't even need any TCP ports anymore, besides for network binding specific tests
1502018-10-02T20:24:05  *** Guyver2 has quit IRC
1512018-10-02T20:24:10  *** rex4539 has joined #bitcoin-core-dev
1522018-10-02T20:41:34  *** timothy has quit IRC
1532018-10-02T21:00:03  *** Chris_Stewart_5 has quit IRC
1542018-10-02T21:10:24  <phantomcircuit> wumpus, indeed that would be an improvement here
1552018-10-02T21:10:29  <phantomcircuit> a pretty significant one actually
1562018-10-02T21:10:33  <phantomcircuit> but yeah priorities
1572018-10-02T21:11:34  <sipa> achow101: hmm, we accept base64 with the == at the end missing?
1582018-10-02T21:22:12  <gwillen> achow101: yes, you can use it as a test vector, it's tesnet with coins from some random faucet and freshly-generated addresses
1592018-10-02T21:22:36  <gwillen> sipa: it seems so, I think it would make sense to not
1602018-10-02T21:22:51  <gwillen> in which case there are two separate issues here
1612018-10-02T21:23:56  <sipa> gwillen: agree; the RFC cited says padding is required; i tgink we should require it
1622018-10-02T21:37:06  *** owowo has quit IRC
1632018-10-02T21:37:54  <gwillen> hm, is there someone in here who is familiar with the QT stuff and QT Creator / QT Designer? Particularly on a mac?
1642018-10-02T21:38:39  <gwillen> I followed the intstructions in the OS X build docs for opening the project in QT Creator, but by default it was unable to see a bunch of headers and I had to add some lines to bitcoin-qt.include to make it avoid drowning everything in red squiggly underlines
1652018-10-02T21:41:14  <phantomcircuit> sipa, lots of things accept base64 without the padding
1662018-10-02T21:41:46  *** owowo has joined #bitcoin-core-dev
1672018-10-02T21:42:03  <wumpus> gwillen: I don't think anyone opens the project in qt creator anymore
1682018-10-02T21:42:34  <gwillen> ha, ok, the build instructions still suggest it
1692018-10-02T21:42:42  <gwillen> I assume QT Designer is still necessary to edit the forms
1702018-10-02T21:42:43  <wumpus> (besides for editing the gui forms)
1712018-10-02T21:42:46  <gwillen> but I guess you can use it standalone?
1722018-10-02T21:42:47  <gwillen> *nods*
1732018-10-02T21:42:58  <gwillen> the fact that I need to edit the gui forms is the only reason I bothered
1742018-10-02T21:43:28  <wumpus> correct, you can use qt designer standalone
1752018-10-02T21:45:47  *** promag has joined #bitcoin-core-dev
1762018-10-02T21:50:05  *** rex4539 has quit IRC
1772018-10-02T21:52:18  <sipa> phantomcircuit: there are multiple base64, some with padding, some without, some mandatory; i wouldn't mind having picked another spec,but the BIP chose one, so we should stick to it
1782018-10-02T21:52:28  *** dqx has quit IRC
1792018-10-02T21:53:22  <phantomcircuit> sipa, ah interesting
1802018-10-02T21:53:34  <phantomcircuit> yeah it doesn't really matter to me and that seems logical
1812018-10-02T21:54:52  *** Chris_Stewart_5 has joined #bitcoin-core-dev
1822018-10-02T22:00:10  <phantomcircuit> lol 0.16.3 installed on windows is stuck at "done loading"
1832018-10-02T22:03:30  <sipa> hmm, rfc 3548 does not require padding at decode time, only when encoding
1842018-10-02T22:18:55  *** hebasto has quit IRC
1852018-10-02T22:19:39  *** Chris_Stewart_5 has quit IRC
1862018-10-02T22:41:59  *** michaelsdunn1 has quit IRC
1872018-10-02T22:55:37  *** promag has quit IRC
1882018-10-02T23:02:28  *** reallll has joined #bitcoin-core-dev
1892018-10-02T23:05:53  *** belcher_ has quit IRC
1902018-10-02T23:11:13  <phantomcircuit> wumpus, like i cant explain how the last commit on #14336 would fix the previous ci failures
1912018-10-02T23:11:15  <gribble> https://github.com/bitcoin/bitcoin/issues/14336 | net: implement poll by pstratem · Pull Request #14336 · bitcoin/bitcoin · GitHub
1922018-10-02T23:14:07  *** Zenton has quit IRC
1932018-10-02T23:14:15  *** Zenton has joined #bitcoin-core-dev
1942018-10-02T23:39:17  *** wullon has joined #bitcoin-core-dev
1952018-10-02T23:42:05  *** promag has joined #bitcoin-core-dev
1962018-10-02T23:45:11  *** drexl has quit IRC
1972018-10-02T23:53:27  *** esotericnonsense has quit IRC
1982018-10-02T23:53:52  *** esotericnonsense has joined #bitcoin-core-dev
1992018-10-02T23:55:51  *** Zenton has quit IRC
2002018-10-02T23:58:43  *** Zenton has joined #bitcoin-core-dev