12016-01-14T00:17:57  *** guest21333 has joined #bitcoin-core-dev
  22016-01-14T00:28:17  <GitHub196> [bitcoin] luke-jr opened pull request #7340: Replace univalue subtree with proper dependency on external UniValue (master...sys_univalue) https://github.com/bitcoin/bitcoin/pull/7340
  32016-01-14T00:33:12  *** jl2012 has quit IRC
  42016-01-14T00:33:32  *** jl2012 has joined #bitcoin-core-dev
  52016-01-14T01:00:49  *** laurentmt has joined #bitcoin-core-dev
  62016-01-14T01:09:39  *** laurentmt has quit IRC
  72016-01-14T01:34:30  *** xiangfu has joined #bitcoin-core-dev
  82016-01-14T01:39:36  *** jtimon has quit IRC
  92016-01-14T01:54:54  *** kangx has joined #bitcoin-core-dev
 102016-01-14T01:55:14  <kangx> hello.
 112016-01-14T02:06:13  *** kangx has quit IRC
 122016-01-14T02:08:07  *** brg444 has quit IRC
 132016-01-14T02:08:21  *** Ylbam has quit IRC
 142016-01-14T02:16:32  *** p15 has joined #bitcoin-core-dev
 152016-01-14T02:19:13  *** adam3us has quit IRC
 162016-01-14T02:49:34  *** adam3us has joined #bitcoin-core-dev
 172016-01-14T02:50:52  *** PaulCape_ has joined #bitcoin-core-dev
 182016-01-14T02:52:57  *** JackH has quit IRC
 192016-01-14T02:53:15  *** PaulCapestany has quit IRC
 202016-01-14T02:53:15  *** Anduck has quit IRC
 212016-01-14T02:53:27  *** Anduck has joined #bitcoin-core-dev
 222016-01-14T02:53:27  *** JackH has joined #bitcoin-core-dev
 232016-01-14T02:59:13  *** adam3us has quit IRC
 242016-01-14T03:01:38  <michagogo> cfields: detached sigs on the way?
 252016-01-14T03:01:54  <cfields> michagogo: signing this very second
 262016-01-14T03:02:23  <cfields> there's a snag with one of the descriptors, but nothing to worry about for rc1 i think
 272016-01-14T03:02:47  <michagogo> Cool, I'll see if I can kick off my script before I leave
 282016-01-14T03:03:36  <cfields> michagogo: you'll have to hand-edit one of the descriptors
 292016-01-14T03:03:44  <michagogo> Hm?
 302016-01-14T03:03:58  <michagogo> How so?
 312016-01-14T03:04:18  <cfields> i'll PR it in a sec, doing sanity checks on the build atm
 322016-01-14T03:04:24  <michagogo> (They're hashed, right?
 332016-01-14T03:04:26  <michagogo> )
 342016-01-14T03:04:32  <cfields> need to remove libc6:i386 out of the osx signer
 352016-01-14T03:05:38  <michagogo> Can you just push it to the 0.12 branch, and then I'll check that out instead of the tag?
 362016-01-14T03:06:31  <cfields> still need to build from the tag, just change the descriptor locally for the sig attach
 372016-01-14T03:06:40  *** brg444 has joined #bitcoin-core-dev
 382016-01-14T03:07:00  <michagogo> But gitian keeps its own git repo and brings in the tag on its own
 392016-01-14T03:07:21  <cfields> yes, but the descriptor is local
 402016-01-14T03:07:59  <michagogo> So if that's the only change to the descriptors from the tag, the clone of the repo that I have outside gitian can be checked out to the branch
 412016-01-14T03:08:08  *** adam3us has joined #bitcoin-core-dev
 422016-01-14T03:08:10  <michagogo> Er, can have the branch checked out
 432016-01-14T03:08:27  <michagogo> Since its ignored other than the descriptors
 442016-01-14T03:08:54  <cfields> you can copy the descriptor to /tmp, it doesn't matter where it lives or what's checked out
 452016-01-14T03:10:06  <michagogo> cfields: I know that. I'm saying, it's probably easiest if you just commit the fix to the 0.12 branch, since I assume that needs to be in rc2 anyway.
 462016-01-14T03:10:27  <cfields> oh
 472016-01-14T03:10:35  <cfields> <cfields> i'll PR it in a sec, doing sanity checks on the build atm
 482016-01-14T03:10:36  <cfields> :)
 492016-01-14T03:10:40  <michagogo> That way the fixed descriptor can easily be used by checking out 0.12 in the external copy of the repo
 502016-01-14T03:11:21  <michagogo> Yeah, just saying that makes the hand-editing unneeded as soon as that's in
 512016-01-14T03:11:46  <cfields> sure, gotcha
 522016-01-14T03:18:53  *** adam3us has quit IRC
 532016-01-14T03:19:02  *** davec has quit IRC
 542016-01-14T03:19:29  *** davec has joined #bitcoin-core-dev
 552016-01-14T03:21:43  <GitHub160> [bitcoin] theuni opened pull request #7342: release: remove libc6 dependency from the osx signing descriptor (master...gitian-quickfix) https://github.com/bitcoin/bitcoin/pull/7342
 562016-01-14T03:23:22  <cfields> sigs pushed
 572016-01-14T03:24:54  *** p15 has quit IRC
 582016-01-14T03:31:07  *** adam3us has joined #bitcoin-core-dev
 592016-01-14T03:43:50  *** thrasher` has left #bitcoin-core-dev
 602016-01-14T03:50:17  *** Squidicuz has joined #bitcoin-core-dev
 612016-01-14T03:52:42  *** jgarzik has joined #bitcoin-core-dev
 622016-01-14T03:53:02  <Luke-Jr> jgarzik: can you 1.0.2 univalue?
 632016-01-14T03:53:40  <Luke-Jr> although it might have a bug, so maybe we should figure that out first..
 642016-01-14T03:54:46  <jgarzik> Luke-Jr, sure
 652016-01-14T03:55:19  <Luke-Jr> jgarzik: any idea on how to determine the cause of https://travis-ci.org/bitcoin/bitcoin/jobs/102242679 ?
 662016-01-14T03:55:23  <Luke-Jr> I can't reproduce it.. :/
 672016-01-14T04:00:35  <cfields> Luke-Jr: it's a string being constructed from setNumStr(), i'd guess null pointer as input
 682016-01-14T04:00:49  <Luke-Jr> hmm
 692016-01-14T04:01:54  <Luke-Jr> if I hack the travis config to --enable-debug, think it'd give line numbers? <.<
 702016-01-14T04:03:16  <cfields> Luke-Jr: you can track it back with that, sec
 712016-01-14T04:03:45  *** PaulCape_ has quit IRC
 722016-01-14T04:04:13  *** PaulCapestany has joined #bitcoin-core-dev
 732016-01-14T04:04:31  <cfields> Luke-Jr: looks like it's internal to univalue, barring some race condition
 742016-01-14T04:04:55  *** ryitpm has quit IRC
 752016-01-14T04:05:20  <Luke-Jr> :x
 762016-01-14T04:05:33  <jgarzik> Luke-Jr, quickest debug solution is running it under gdb in a VM that mimics travis OS setup
 772016-01-14T04:05:43  *** Chris_Stewart_5 has quit IRC
 782016-01-14T04:05:47  <jgarzik> Smart to have such a VM lying around anyway
 792016-01-14T04:05:55  <Luke-Jr> that doesn't sound quick, considering Travis is closed source.. :/
 802016-01-14T04:06:02  <cfields> UniValue::setInt(int64_t val) -> setNumStr()
 812016-01-14T04:06:13  <jgarzik> the OS setup is easy to reproduce
 822016-01-14T04:06:21  <cfields> looks like your crash is at univalue.cpp:129
 832016-01-14T04:06:57  <jgarzik> setNumStr(oss.str());
 842016-01-14T04:07:15  <cfields> Luke-Jr: throw an assert(!oss.empty()) in there
 852016-01-14T04:07:59  <jgarzik> if css is empty, something is majorly broken with libstdc++
 862016-01-14T04:08:07  * jgarzik kicks autocorrect
 872016-01-14T04:08:14  <cfields> no clue how that would happen though
 882016-01-14T04:08:34  *** ryitpm has joined #bitcoin-core-dev
 892016-01-14T04:08:58  <cfields> jgarzik: yea, threading/shutdown issue seems much more likely
 902016-01-14T04:10:18  *** nanotube has quit IRC
 912016-01-14T04:12:26  <Luke-Jr> ⁂ glibc detected ⁂ /home/travis/build/bitcoin/bitcoin/bitcoin-i686-pc-linux-gnu/src/bitcoin-cli: free(): invalid pointer: 0xb7774348 ⁂
 922016-01-14T04:12:32  <Luke-Jr> really doesn't suggest a NULL ptr thing
 932016-01-14T04:12:35  <Luke-Jr> IMO
 942016-01-14T04:13:46  <cfields> Luke-Jr: yea
 952016-01-14T04:14:45  *** nanotube has joined #bitcoin-core-dev
 962016-01-14T04:15:33  *** adam3us has quit IRC
 972016-01-14T04:21:35  <Luke-Jr> 1 hour download for old Ubuntu Server that Travis uses.. :/
 982016-01-14T04:25:34  * Luke-Jr kicks DSL
 992016-01-14T04:25:41  <Luke-Jr> at least I've confirmed Travis reproduces it..
1002016-01-14T04:35:39  <cfields> Luke-Jr: looking more closely, i think it's probably at univalue.cpp:118
1012016-01-14T04:35:58  <cfields> new guess is that the object is already destructed when that assign is attempted
1022016-01-14T04:38:52  *** jcorgan|away is now known as jcorgan
1032016-01-14T04:50:19  <cfields> Luke-Jr: bitcoin-cli.cpp:214 as an example. Isn't that returning a bound temporary?
1042016-01-14T04:52:08  <Luke-Jr>     const UniValue& reply = valReply.get_obj(); ?
1052016-01-14T04:52:30  <Luke-Jr> seems get_obj just throws or returns *this
1062016-01-14T04:55:16  <cfields> Luke-Jr: yes, but it's a temporary bound to valReply, which goes out of scope when the function returns.
1072016-01-14T04:55:36  <Luke-Jr> oh, right
1082016-01-14T04:55:43  <Luke-Jr> didn't realise it was being returned
1092016-01-14T04:55:55  <cfields> unless there's some funky RVO that makes that legal
1102016-01-14T04:56:08  <Luke-Jr> doubt it
1112016-01-14T04:56:41  <Luke-Jr> oh
1122016-01-14T04:56:50  <Luke-Jr> it's probably because the return type is not a reference
1132016-01-14T04:56:52  <Luke-Jr> so it makes a copy
1142016-01-14T04:56:56  <Luke-Jr> ?
1152016-01-14T04:57:13  <jgarzik> correct - return value is not a reference
1162016-01-14T04:58:12  <jgarzik> FYI - get_obj() is legacy and should not be used
1172016-01-14T04:58:33  <jgarzik> it's just a dumb type check with a nutty assignment style
1182016-01-14T04:58:35  *** guest21333 has quit IRC
1192016-01-14T04:59:33  <jgarzik> i.e. the purpose is simply to throw an exception if the type check fails
1202016-01-14T05:01:36  <cfields> erm, that still doesn't look right
1212016-01-14T05:06:59  <cfields> http://en.cppreference.com/w/cpp/language/reference_initialization
1222016-01-14T05:07:11  <cfields> see the first exception under "Lifetime of a temporary"
1232016-01-14T05:10:25  *** blkdb has quit IRC
1242016-01-14T05:10:36  *** blkdb has joined #bitcoin-core-dev
1252016-01-14T05:11:43  <Luke-Jr> cfields: but it's not a temporary, it's scoped
1262016-01-14T05:12:16  <Luke-Jr> reply and valReply are literally the same thing
1272016-01-14T05:14:20  <Luke-Jr> lovely, ubuntu-12.04.5-server-amd64.iso doesn't work in KVM
1282016-01-14T05:14:29  <Luke-Jr> (and probably not at all)
1292016-01-14T05:21:27  <cfields> mm, ok
1302016-01-14T05:34:18  <Luke-Jr> put bitcoin-cli through valgrind for that rpc-test and it sees no issues
1312016-01-14T05:38:04  *** xiangfu has quit IRC
1322016-01-14T05:51:42  *** xiangfu has joined #bitcoin-core-dev
1332016-01-14T06:01:28  *** adam3us has joined #bitcoin-core-dev
1342016-01-14T06:22:24  *** xiangfu has quit IRC
1352016-01-14T06:22:41  *** xiangfu has joined #bitcoin-core-dev
1362016-01-14T06:26:29  *** p15 has joined #bitcoin-core-dev
1372016-01-14T06:42:52  *** brg444 has quit IRC
1382016-01-14T06:53:41  *** Ylbam has joined #bitcoin-core-dev
1392016-01-14T07:12:58  *** jcorgan is now known as jcorgan|away
1402016-01-14T08:13:18  <jonasschnelli> Luke-Jr: re: removing univalue subtree:
1412016-01-14T08:13:45  <jonasschnelli> I guess UniValue is not available over pkg managers, so people need to install the dependency by checking out the git repo and compile?
1422016-01-14T08:15:39  *** jtimon has joined #bitcoin-core-dev
1432016-01-14T08:16:23  *** guest21333 has joined #bitcoin-core-dev
1442016-01-14T08:19:38  *** paveljanik has joined #bitcoin-core-dev
1452016-01-14T08:19:38  *** paveljanik has joined #bitcoin-core-dev
1462016-01-14T08:33:12  *** guest21333 has quit IRC
1472016-01-14T08:44:44  *** jtimon has quit IRC
1482016-01-14T08:46:05  *** BashCo has quit IRC
1492016-01-14T09:07:38  *** BashCo has joined #bitcoin-core-dev
1502016-01-14T09:09:24  *** moli has quit IRC
1512016-01-14T09:14:59  *** zookolaptop has quit IRC
1522016-01-14T09:43:55  *** xiangfu has quit IRC
1532016-01-14T09:44:11  *** xiangfu has joined #bitcoin-core-dev
1542016-01-14T09:51:06  *** randy-waterhouse has quit IRC
1552016-01-14T10:02:04  *** Guyver2 has joined #bitcoin-core-dev
1562016-01-14T10:39:19  *** xiangfu has quit IRC
1572016-01-14T10:49:46  *** Guyver2 has quit IRC
1582016-01-14T11:26:35  *** Thireus has joined #bitcoin-core-dev
1592016-01-14T11:58:26  *** Naphex has quit IRC
1602016-01-14T12:01:47  *** laurentmt has joined #bitcoin-core-dev
1612016-01-14T12:03:09  *** laurentmt has quit IRC
1622016-01-14T12:09:08  *** jtimon has joined #bitcoin-core-dev
1632016-01-14T12:46:37  <GitHub176> [bitcoin] laanwj pushed 2 new commits to master: https://github.com/bitcoin/bitcoin/compare/c079d79c9a9c...2be4bb51f37c
1642016-01-14T12:46:38  <GitHub176> bitcoin/master 3503a78 Cory Fields: release: remove libc6 dependency from the osx signing descriptor...
1652016-01-14T12:46:38  <GitHub176> bitcoin/master 2be4bb5 Wladimir J. van der Laan: Merge pull request #7342...
1662016-01-14T12:46:48  <GitHub199> [bitcoin] laanwj closed pull request #7342: release: remove libc6 dependency from the osx signing descriptor (master...gitian-quickfix) https://github.com/bitcoin/bitcoin/pull/7342
1672016-01-14T12:47:13  <GitHub96> [bitcoin] laanwj pushed 1 new commit to 0.12: https://github.com/bitcoin/bitcoin/commit/fbea2f64e23487636f2e490b6b0067c41545cd39
1682016-01-14T12:47:13  <GitHub96> bitcoin/0.12 fbea2f6 Cory Fields: release: remove libc6 dependency from the osx signing descriptor...
1692016-01-14T12:56:52  <GitHub64> [bitcoin] laanwj pushed 2 new commits to master: https://github.com/bitcoin/bitcoin/compare/2be4bb51f37c...e1060c56cca7
1702016-01-14T12:56:52  <GitHub64> bitcoin/master fa989fb MarcoFalke: [qt] coincontrol workaround is still needed in qt5.4 (fixed in qt5.5)
1712016-01-14T12:56:53  <GitHub64> bitcoin/master e1060c5 Wladimir J. van der Laan: Merge pull request #7334...
1722016-01-14T12:57:02  <GitHub186> [bitcoin] laanwj closed pull request #7334: [qt] coincontrol workaround is still needed in qt5.4 (fixed in qt5.5) (master...Mf1601-qt55Workaround) https://github.com/bitcoin/bitcoin/pull/7334
1732016-01-14T13:02:26  *** berndj has quit IRC
1742016-01-14T13:07:55  *** laurentmt has joined #bitcoin-core-dev
1752016-01-14T13:10:46  *** laurentmt has quit IRC
1762016-01-14T13:15:03  *** berndj has joined #bitcoin-core-dev
1772016-01-14T13:15:46  *** berndj has quit IRC
1782016-01-14T13:16:01  *** berndj has joined #bitcoin-core-dev
1792016-01-14T13:21:40  *** Thireus has quit IRC
1802016-01-14T13:58:54  *** Chris_Stewart_5 has joined #bitcoin-core-dev
1812016-01-14T14:01:56  *** jtimon has quit IRC
1822016-01-14T14:02:33  *** jonasschnelli has quit IRC
1832016-01-14T14:10:21  *** jonasschnelli has joined #bitcoin-core-dev
1842016-01-14T14:23:42  *** p15 has quit IRC
1852016-01-14T14:39:18  *** Chris_Stewart_5 has quit IRC
1862016-01-14T14:41:17  *** laurentmt has joined #bitcoin-core-dev
1872016-01-14T14:47:57  <sdaftuar> i was just looking over the release notes for 0.12 -- am i right in thinking we should highlight chainstate obfuscation and discuss downgrade issues that coudl result?
1882016-01-14T14:48:46  <sdaftuar> (i'm not that familiar with the topic but vaguely recall people talking about it before)
1892016-01-14T14:52:22  *** Chris_Stewart_5 has joined #bitcoin-core-dev
1902016-01-14T14:54:10  <Luke-Jr> jonasschnelli: or depends/ would presumably also work fine
1912016-01-14T14:54:20  <Luke-Jr> sdaftuar: agree
1922016-01-14T14:54:52  <jonasschnelli> Luke-Jr: you mean for univalue?
1932016-01-14T14:55:09  <Luke-Jr> jonasschnelli: yes, or any other dependency
1942016-01-14T14:55:24  <jonasschnelli> Yes. But it would be the first dependency that would require self-compiling.
1952016-01-14T14:55:47  <Luke-Jr> only until package managers pick it up *shrug*
1962016-01-14T14:56:00  <Luke-Jr> and only in the case of the user compiling in the first place :p
1972016-01-14T15:00:53  *** laurentmt has quit IRC
1982016-01-14T15:16:20  <Luke-Jr> cfields: test_bitcoin: key.cpp:305: void ECC_Start(): Assertion `secp256k1_context_sign == __null' failed.
1992016-01-14T15:16:22  <Luke-Jr> O.o
2002016-01-14T15:40:30  <jgarzik> Luke-Jr, Bloq (my newco) will be submitting Fedora and Debian packages upstream for univalue
2012016-01-14T15:40:41  <jgarzik> that's a long process though, maybe a year
2022016-01-14T15:40:52  <jgarzik> before it trickles down to a stable release
2032016-01-14T15:40:58  <Luke-Jr> jgarzik: cool
2042016-01-14T15:41:35  <jgarzik> It's not consensus critical so it makes sense as a package rather than subtree in the long term, IMO
2052016-01-14T15:42:25  <Luke-Jr> yeah, and we have a package manager bundled anyway
2062016-01-14T15:43:51  <Luke-Jr> in the meantime, here's a full stack trace with line numbers for the free bug http://codepad.org/bKXLXqia
2072016-01-14T15:45:40  <jgarzik> Luke-Jr, this is for a tree that is up to date with jgarzik/univalue master?
2082016-01-14T15:45:46  <Luke-Jr> no, this is 1.0.1
2092016-01-14T15:49:26  <jgarzik> Luke-Jr, any chance with http://codepad.org/u89EfIeE ?
2102016-01-14T15:49:31  <jgarzik> *change
2112016-01-14T15:51:25  <Luke-Jr> I don't have an easy way to test changes in univalue right now, so it will be a bit to answer
2122016-01-14T16:04:44  *** jtimon has joined #bitcoin-core-dev
2132016-01-14T16:05:09  <jonasschnelli> I'm fine with keeping it at jgarzik/univalue ... I think we have moved it because we wanted to get a PR merged? not sure although.
2142016-01-14T16:05:31  <jonasschnelli> jgarzik: nice that you try to get upstream packages!
2152016-01-14T16:09:25  <jgarzik> jonasschnelli, basic open source upstreaming policy :)
2162016-01-14T16:10:24  <Luke-Jr> #10 0xf7fca7ca in UniValue::setNumStr (this=0xffffd348, val_=…) at lib/univalue.cpp:118
2172016-01-14T16:10:25  <Luke-Jr> 118         val = val_;
2182016-01-14T16:11:22  <Luke-Jr> prior to this, val is:
2192016-01-14T16:11:24  <Luke-Jr> (gdb) print val
2202016-01-14T16:11:26  <Luke-Jr> $4 = {static npos = <optimized out>, _M_dataplus = {<std::allocator<char>> = {<__gnu_cxx::new_allocator<char>> = {<No data fields>}, <No data fields>},
2212016-01-14T16:11:27  <Luke-Jr>     _M_p = 0x567d93d4 ""}}
2222016-01-14T16:12:27  <jonasschnelli> ^^
2232016-01-14T16:13:15  <Luke-Jr> in this case the error is:
2242016-01-14T16:13:17  <Luke-Jr> ⁂ glibc detected ⁂ /home/ubuntu/bitcoin/src/bitcoin-cli: free(): invalid pointer: 0x567d93c8 ⁂
2252016-01-14T16:18:59  *** laurentmt has joined #bitcoin-core-dev
2262016-01-14T16:23:27  *** laurentmt has quit IRC
2272016-01-14T16:27:00  *** SmiteMeSmith__ has quit IRC
2282016-01-14T16:32:01  <cfields> Luke-Jr: i see that pretty often, iirc it happens if something goes wrong during startup
2292016-01-14T16:45:06  *** SmiteMeSmith__ has joined #bitcoin-core-dev
2302016-01-14T17:09:00  *** BashCo has quit IRC
2312016-01-14T17:12:52  *** ryitpm has quit IRC
2322016-01-14T17:15:11  *** ryitpm has joined #bitcoin-core-dev
2332016-01-14T17:18:31  *** zookolaptop has joined #bitcoin-core-dev
2342016-01-14T17:20:16  *** zookolaptop has quit IRC
2352016-01-14T17:22:43  *** zookolaptop has joined #bitcoin-core-dev
2362016-01-14T17:23:41  *** zookolaptop is now known as zooko
2372016-01-14T17:23:46  *** jtimon has quit IRC
2382016-01-14T17:25:59  <cfields> Luke-Jr: mind doing a quick check on #7322 ?
2392016-01-14T17:27:28  *** zooko` has joined #bitcoin-core-dev
2402016-01-14T17:27:31  *** zooko has quit IRC
2412016-01-14T17:28:00  <Luke-Jr> testing
2422016-01-14T17:28:22  *** zooko` is now known as zooko
2432016-01-14T17:30:04  <cfields> thanks
2442016-01-14T17:30:05  *** zooko has quit IRC
2452016-01-14T17:32:33  *** zooko has joined #bitcoin-core-dev
2462016-01-14T17:34:59  <cfields> <sdaftuar> i was just looking over the release notes for 0.12 -- am i right in thinking we should highlight chainstate obfuscation and discuss downgrade issues that coudl result?
2472016-01-14T17:35:04  <cfields> ^^ yes
2482016-01-14T17:35:30  <cfields> i accidentally ran an obfuscated chainstate on an older version, didn't go well
2492016-01-14T17:38:26  <Luke-Jr> it didn't just immediately fail? :x
2502016-01-14T17:39:00  <cfields> Luke-Jr: of course, but to a casual user the "why" wouldn't be obvious
2512016-01-14T17:39:05  <Luke-Jr> ah
2522016-01-14T17:48:05  *** adam3us has quit IRC
2532016-01-14T17:49:55  *** moli has joined #bitcoin-core-dev
2542016-01-14T17:52:44  *** adam3us has joined #bitcoin-core-dev
2552016-01-14T17:57:26  <cfields> Luke-Jr: aha, i didn't notice the beginning of that stacktrace. facepalm.
2562016-01-14T17:57:36  *** jtimon has joined #bitcoin-core-dev
2572016-01-14T17:59:44  <Luke-Jr> ?
2582016-01-14T18:00:30  <cfields> Luke-Jr: it's gone off the rails already in main()
2592016-01-14T18:01:55  <cfields> Luke-Jr: $1 says if you build univalue static, travis passes
2602016-01-14T18:02:27  <Luke-Jr> cfields: that doesn't fix the problem :p
2612016-01-14T18:02:36  <cfields> Luke-Jr: sure, just trying to narrow it down
2622016-01-14T18:03:22  * Luke-Jr stabs Ubuntu for valgrind not working
2632016-01-14T18:04:17  <cfields> Luke-Jr: btw, you can test on travis much quicker if you delete the other builders in your .travis file
2642016-01-14T18:04:26  <cfields> i've resorted to that once or twice
2652016-01-14T18:05:28  <Luke-Jr> I have it reproduced in a gitian VM :P
2662016-01-14T18:05:41  <cfields> oh, great
2672016-01-14T18:07:19  <Luke-Jr> making univalue static in bitcoin-cli results in bitcoind crashing instead
2682016-01-14T18:07:26  *** brg444 has joined #bitcoin-core-dev
2692016-01-14T18:07:26  <Luke-Jr> so you're probably right
2702016-01-14T18:08:11  <cfields> init order issue?
2712016-01-14T18:09:58  *** MarcoFalke has joined #bitcoin-core-dev
2722016-01-14T18:15:22  <cfields> Luke-Jr: maybe there's an issue with cmdline args being parsed via univalue at startup? do you know what the passed-in args are?
2732016-01-14T18:18:29  <Luke-Jr> it doesn't seem to matter
2742016-01-14T18:19:19  <cfields> any args cause a crash?
2752016-01-14T18:20:28  <Luke-Jr> no, it needs at least rpcpassword and a method
2762016-01-14T18:20:29  *** Thireus has joined #bitcoin-core-dev
2772016-01-14T18:25:03  <Luke-Jr> cfields: http://paste.ubuntu.com/14497754/ fixes it. :|
2782016-01-14T18:25:47  <cfields> wha?
2792016-01-14T18:26:46  <cfields> so something's assuming that it has a string value without checking it, i suppose?
2802016-01-14T18:27:30  <Luke-Jr> I was given this idea from https://gcc.gnu.org/bugzilla/show_bug.cgi?id=31368
2812016-01-14T18:29:45  <Luke-Jr> using : val() is a crash again
2822016-01-14T18:32:49  <GitHub11> [bitcoin] MarcoFalke opened pull request #7345: [0.12]  	Add backwards-compatibility issues (chainstate obfuscation) to release notes (0.12...Mf1601-docObfuscation) https://github.com/bitcoin/bitcoin/pull/7345
2832016-01-14T18:37:16  <Luke-Jr> cfields: : val("") also crashes
2842016-01-14T18:39:48  *** maaku has joined #bitcoin-core-dev
2852016-01-14T18:39:55  *** maaku has left #bitcoin-core-dev
2862016-01-14T18:41:05  <cfields> Luke-Jr: can you build with -O0 and do a 'bt full' in gdb?
2872016-01-14T18:41:09  <cfields> i'm really curious now :\
2882016-01-14T18:42:08  <Luke-Jr> http://codepad.org/P3a2sD7h
2892016-01-14T18:43:37  <Luke-Jr> cfields: moving the constructor definition to the lib rather than inline, also seems to work
2902016-01-14T18:43:50  <cfields> Luke-Jr: yes, i was just about to suggest that
2912016-01-14T18:44:09  <cfields> Luke-Jr: i'm afraid it has to do with compile differences or locales
2922016-01-14T18:44:23  <Luke-Jr> ?
2932016-01-14T18:45:13  <Luke-Jr> crash remains with --disable-glibc-back-compat FWIW
2942016-01-14T18:45:45  <cfields> oh wait
2952016-01-14T18:59:11  <Luke-Jr> ?
2962016-01-14T19:00:15  *** kangx has joined #bitcoin-core-dev
2972016-01-14T19:00:17  <cfields> i'm re-reading the _GLIBCXX_DEBUG docs, trying to see if there are issues passing objects in/out of shared libs
2982016-01-14T19:00:26  *** devrando2 has joined #bitcoin-core-dev
2992016-01-14T19:01:13  <cfields> Luke-Jr: can you verify that bitcoin-cli.cpp is built with those flags?
3002016-01-14T19:01:27  <Luke-Jr> which flags?
3012016-01-14T19:01:56  <btcdrak> meeting in #bitcoin-dev now
3022016-01-14T19:06:43  *** treehug88 has joined #bitcoin-core-dev
3032016-01-14T19:13:31  *** Thireus has quit IRC
3042016-01-14T19:33:33  <MarcoFalke> :( Looks like travis doesn't like the nightly builds. Can someone after the meeting check https://blog.travis-ci.com/2014-05-12-why-is-my-build-not-running/
3052016-01-14T19:33:52  <MarcoFalke> nightly is running for 2:20 h. (2 h is the max, iirc)
3062016-01-14T19:35:12  <GitHub142> [bitcoin] luke-jr opened pull request #7346: 0.12 release notes: Mining Policy Changes (0.12...0.12-release-notes-mining) https://github.com/bitcoin/bitcoin/pull/7346
3072016-01-14T19:36:47  <Luke-Jr> cfields: what flags?
3082016-01-14T19:47:10  *** brg444 has quit IRC
3092016-01-14T19:50:40  *** BashCo has joined #bitcoin-core-dev
3102016-01-14T19:53:58  *** tulip has joined #bitcoin-core-dev
3112016-01-14T19:56:38  *** tulip has left #bitcoin-core-dev
3122016-01-14T19:57:47  *** zooko` has joined #bitcoin-core-dev
3132016-01-14T19:59:29  *** zooko has quit IRC
3142016-01-14T20:13:05  <GitHub181> [bitcoin] morcos opened pull request #7347: Release notes about mining code changes (0.12...rnfixes2) https://github.com/bitcoin/bitcoin/pull/7347
3152016-01-14T20:13:23  *** adam3us has quit IRC
3162016-01-14T20:17:08  *** brg444 has joined #bitcoin-core-dev
3172016-01-14T20:19:39  <Luke-Jr> morcos: what's this nonsense about removal of priority code?
3182016-01-14T20:23:39  <Luke-Jr> I guess it predates your PR
3192016-01-14T20:23:59  * Luke-Jr fixes in #7346
3202016-01-14T20:24:19  <Luke-Jr> morcos: in any case, please rebase 7347 on top of that
3212016-01-14T20:26:46  *** devrando2 has quit IRC
3222016-01-14T20:27:38  <morcos> Luke-Jr: I did not add that.  I reworded what already was in the release notes that were merged.  I think if you read my changes, you'll agree they did not change meaning.
3232016-01-14T20:28:14  <Luke-Jr> morcos: yes, noticed that [20:23:38] <Luke-Jr> I guess it predates your PR
3242016-01-14T20:28:26  <morcos> oh, sorry, skipped over that
3252016-01-14T20:29:54  <Luke-Jr> sorry for having jumped to conclusions at first
3262016-01-14T20:30:11  *** zooko` is now known as zooko
3272016-01-14T20:31:18  <morcos> Well as you know I think that is what we should do, but I wouldn't have suggested to put it in there without clearly pointing it out, as I know there is not complete consensus on it
3282016-01-14T20:38:09  *** arowser has quit IRC
3292016-01-14T20:38:26  *** Guyver2 has joined #bitcoin-core-dev
3302016-01-14T20:38:32  *** zooko has quit IRC
3312016-01-14T20:38:36  *** arowser has joined #bitcoin-core-dev
3322016-01-14T20:39:16  <Luke-Jr> sigh.. reddit.. downvoted for telling people not to use a literally random yum repository for Bitcoin Core ..
3332016-01-14T20:42:56  <paveljanik> and the answer was do not trust the source, read it first? ;-)
3342016-01-14T21:00:15  *** MarcoFalke has quit IRC
3352016-01-14T21:01:17  *** MarcoFalke has joined #bitcoin-core-dev
3362016-01-14T21:15:42  *** adam3us has joined #bitcoin-core-dev
3372016-01-14T21:21:49  *** BashCo has quit IRC
3382016-01-14T21:22:30  *** BashCo has joined #bitcoin-core-dev
3392016-01-14T21:22:40  *** murch has joined #bitcoin-core-dev
3402016-01-14T21:51:04  *** [1]evoskuil has joined #bitcoin-core-dev
3412016-01-14T21:53:26  *** evoskuil has quit IRC
3422016-01-14T21:53:26  *** [1]evoskuil is now known as evoskuil
3432016-01-14T22:01:12  *** treehug88 has quit IRC
3442016-01-14T22:08:33  *** brg444_ has joined #bitcoin-core-dev
3452016-01-14T22:09:29  *** brg444_ has quit IRC
3462016-01-14T22:10:10  *** brg444 has quit IRC
3472016-01-14T22:11:17  *** brg444 has joined #bitcoin-core-dev
3482016-01-14T22:24:11  *** MarcoFalke has quit IRC
3492016-01-14T22:36:29  *** Thireus has joined #bitcoin-core-dev
3502016-01-14T22:52:52  *** Guyver2 has quit IRC
3512016-01-14T23:15:18  *** murch has quit IRC
3522016-01-14T23:15:43  *** murch has joined #bitcoin-core-dev
3532016-01-14T23:15:48  *** zookolaptop has joined #bitcoin-core-dev
3542016-01-14T23:33:02  <PRab> I'm getting "cannot stat ‘base-trusty-amd64’: No such file or directory" when trying to run gbuild. Any ideas?
3552016-01-14T23:33:25  <Luke-Jr> cfields: http://paste.ubuntu.com/14499832/ also fixes the problem
3562016-01-14T23:33:34  <Luke-Jr> PRab: you need to build a trusty VM base
3572016-01-14T23:34:21  <PRab> Luke-Jr: How? I don't see that mentioned in https://github.com/bitcoin/bitcoin/blob/master/doc/release-process.md#perform-gitian-builds
3582016-01-14T23:34:47  <cfields> Luke-Jr: so it's being instantiated two different ways?
3592016-01-14T23:35:07  <Luke-Jr> cfields: no, it seems to be a libstdc++ bug https://gcc.gnu.org/ml/libstdc++/2006-01/msg00162.html
3602016-01-14T23:36:08  <cfields> Luke-Jr: we don't use a custom string type, though?
3612016-01-14T23:36:35  <Luke-Jr> cfields: it doesn't seem to be restricted to only custom ones :/
3622016-01-14T23:36:44  <Luke-Jr> at least in this old version
3632016-01-14T23:36:58  <Luke-Jr> PRab: https://github.com/bitcoin/bitcoin/blob/master/doc/gitian-building.md#installing-gitian
3642016-01-14T23:37:04  *** brg444 has quit IRC
3652016-01-14T23:37:40  *** zookolaptop has quit IRC
3662016-01-14T23:38:38  <Luke-Jr> cfields: maybe bitcoin-cli is so simple that it doesn't instantise std::string sufficiently on its own
3672016-01-14T23:38:49  <cfields> Luke-Jr: i still suspect that it has to do with the STDCXX debug option on a shared lib, when compiled differently
3682016-01-14T23:39:10  <PRab> Luke-Jr: Thanks. I didn't realize I would need to redo any of that. In the past I was able to get away with just following the release-process instructions. I'm guessing it used an older version of ubuntu before.
3692016-01-14T23:39:10  <Luke-Jr> I don't understand what that means. I am compiling them the same afaik
3702016-01-14T23:39:22  <Luke-Jr> PRab: yes
3712016-01-14T23:40:02  <Luke-Jr> jgarzik: so it seems this issue is probably unrelated to univalue, so 1.0.2 tag seems good to go IMO?
3722016-01-14T23:40:17  <cfields> Luke-Jr: apparently libstdc++ uses a single address to notate that a string is "empty", and avoids freeing it in that case. If the addresses end up being different, an empty string allocated on the bitcoin side would attempt to be freed by the lib
3732016-01-14T23:41:05  <cfields> which would also explain why avoiding the assignment in the header would "fix" the issue
3742016-01-14T23:41:08  <Luke-Jr> cfields: hmm
3752016-01-14T23:41:26  <Luke-Jr> how is that single address determined? :x
3762016-01-14T23:41:28  <cfields> Luke-Jr: you use -fPIC and our hardening opts to build the lib?
3772016-01-14T23:42:06  <cfields> (and visibility)
3782016-01-14T23:42:29  <Luke-Jr> probably not.. those shouldn't affect ABI though :|
3792016-01-14T23:42:46  <cfields> PIC certainly does
3802016-01-14T23:42:55  <cfields> though visibility would get ugly, since univalue doesn't handle it
3812016-01-14T23:43:03  <Luke-Jr> well, libs are going to be PIC
3822016-01-14T23:43:17  <JackH> hi now that we are in bitcoin-core-dev channel, can I suggest something to the core dev's openly
3832016-01-14T23:43:19  <Luke-Jr> that last pastebin doesn't make sense with your theory, does it?
3842016-01-14T23:44:12  <Luke-Jr> jgarzik: actually, maybe the std::string workaround ought to be in univalue.h?
3852016-01-14T23:44:35  <cfields> Luke-Jr: i must admit that one makes no sense to me
3862016-01-14T23:44:53  <cfields> unless it's just avoiding them being defined differently
3872016-01-14T23:46:15  <cfields> Luke-Jr: i'm going to see if i can reproduce locally
3882016-01-14T23:46:38  <Luke-Jr> cfields: want me to just open my VM to you over the net?
3892016-01-14T23:46:49  <cfields> oh, sure. that'd be even easier :)
3902016-01-14T23:47:13  <cfields> Luke-Jr: not doubting the fix, just trying to understand what's going on
3912016-01-14T23:47:52  *** laurentmt has joined #bitcoin-core-dev
3922016-01-14T23:48:12  *** laurentmt has quit IRC
3932016-01-14T23:53:26  <Luke-Jr> cfields: I'm currently testing the idea of putting that line in the univalue.h fwi
3942016-01-14T23:53:27  <Luke-Jr> fwiw
3952016-01-14T23:54:42  <cfields> Luke-Jr: won't that cause nasty link problems?
3962016-01-14T23:55:09  <Luke-Jr> I guess we'll find out <.<
3972016-01-14T23:55:21  <Luke-Jr> might make the compile much slower though? :/
3982016-01-14T23:55:34  <Luke-Jr> perhaps moving the constructors to the lib is the best option
3992016-01-14T23:56:02  *** Thireus has quit IRC
4002016-01-14T23:56:19  *** kangx has quit IRC
4012016-01-14T23:56:24  <Luke-Jr> certainly feels like this build is slower..
4022016-01-14T23:57:41  <Luke-Jr> cfields: ok, back to vanilla univalue; all yours
4032016-01-14T23:57:54  <Luke-Jr> in case you missed it, the rpc test passed with the template hack
4042016-01-14T23:58:16  <cfields> hmm, ok
4052016-01-14T23:58:58  <Luke-Jr> (Ctrl-A Esc to use pgup/pgdown keys to scroll in screen)
4062016-01-14T23:59:09  <Luke-Jr> (note it freezes the screen until you hit Esc again)