19:00:22 <wumpus> #startmeeting
19:00:22 <lightningbot> Meeting started Thu Feb 14 19:00:22 2019 UTC.  The chair is wumpus. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:22 <lightningbot> Useful Commands: #action #agreed #help #info #idea #link #topic.
19:00:27 <provoostenator> hi
19:00:29 <jamesob> hi
19:00:33 <instagibbs> hi
19:00:38 <wumpus> #bitcoin-core-dev Meeting: wumpus sipa gmaxwell jonasschnelli morcos luke-jr sdaftuar jtimon cfields petertodd kanzure bluematt instagibbs phantomcircuit codeshark michagogo marcofalke paveljanik NicolasDorier jl2012 achow101 meshcollider jnewbery maaku fanquake promag provoostenator aj Chris_Stewart_5 dongcarl gwillen jamesob ken281221 ryanofsky gleb
19:00:39 <moneyball> topic proposed by provoostenator: minimum QT version for v0.18, see #13478
19:00:41 <gribble> https://github.com/bitcoin/bitcoin/issues/13478 | [RFC] gui: Minimum required Qt5 · Issue #13478 · bitcoin/bitcoin · GitHub
19:00:43 <promag> hi
19:00:45 <jonasschnelli> hi
19:00:50 <achow101> hi
19:01:14 <moneyball> happy valentine's day :)
19:01:20 <luke-jr> St. Valentine's day*
19:01:22 <jamesob> <3
19:01:33 <provoostenator> Thanks! I spent it with AppVeyor mostly :-)
19:01:37 <luke-jr> lol
19:01:48 <jonasschnelli> heh
19:02:48 <wumpus> 💜
19:03:16 <luke-jr> Sjors pointed me to the issue when I asked what benefits there are, but I didn't see any there..
19:03:49 <luke-jr> I also don't see where/if anyone done a check for what versions come with current stable distros.
19:03:51 <wumpus> #topic 0.18 feature freeze
19:03:54 <phantomcircuit> hi
19:03:55 <phantomcircuit> luke-jr, happy halmark day
19:04:02 <provoostenator> luke-jr: the issue described the reasons for switching, but those reasons aren't benefits :-)
19:04:30 <sipa> will be there in 5-10 mi
19:04:32 <sipa> min
19:05:05 <wumpus> tomorrow (15th) is the 0.18 feature freeze
19:05:22 <wumpus> there's still quite a few PRs open on the 0.18 milestone https://github.com/bitcoin/bitcoin/pulls?q=is%3Aopen+is%3Apr+milestone%3A0.18.0 though not all are features of course
19:06:54 <wumpus> but it'd definitely help to get some more review there
19:06:56 <gleb> hi
19:07:19 <phantomcircuit> luke-jr, generally 4/5 are available in parallel
19:08:00 <gmaxwell> hi
19:08:05 <luke-jr> phantomcircuit: ?
19:08:14 <phantomcircuit> luke-jr, qt
19:08:23 <phantomcircuit> debian at least i have both installed
19:08:27 <wumpus> still open features are #15195 for the GUI and a few RPC ones
19:08:29 <gribble> https://github.com/bitcoin/bitcoin/issues/15195 | gui: Add Close Wallet action by promag · Pull Request #15195 · bitcoin/bitcoin · GitHub
19:08:30 <luke-jr> phantomcircuit: we already don't support Qt4
19:08:33 <wumpus> please, stick to the topic
19:08:39 <phantomcircuit> oh right
19:09:51 <wumpus> for RPC there's #14481 #14075 #14021 #13932 left that are really features
19:09:53 <gribble> https://github.com/bitcoin/bitcoin/issues/14481 | Add P2SH-P2WSH support to listunspent RPC by MeshCollider · Pull Request #14481 · bitcoin/bitcoin · GitHub
19:09:56 <gribble> https://github.com/bitcoin/bitcoin/issues/14075 | Import watch only pubkeys to the keypool if private keys are disabled by achow101 · Pull Request #14075 · bitcoin/bitcoin · GitHub
19:09:58 <gribble> https://github.com/bitcoin/bitcoin/issues/14021 | Import key origin data through descriptors in importmulti by achow101 · Pull Request #14021 · bitcoin/bitcoin · GitHub
19:10:02 <gribble> https://github.com/bitcoin/bitcoin/issues/13932 | Additional utility RPCs for PSBT by achow101 · Pull Request #13932 · bitcoin/bitcoin · GitHub
19:10:04 <kanzure> hi.
19:10:29 <instagibbs> key origin import seems camera-ready if others want to look
19:11:02 <achow101> 14021 is pretty close
19:11:18 <wumpus> good!
19:12:11 <promag> I'm waiting for some feedback on 15195
19:12:12 <wumpus> ah there's also #15368
19:12:14 <gribble> https://github.com/bitcoin/bitcoin/issues/15368 | Descriptor checksums by sipa · Pull Request #15368 · bitcoin/bitcoin · GitHub
19:12:48 <gmaxwell> luke-jr: do you have a 32-bit host with rdrand (modern intel/amd cpu running in 32bit mode)? would be nice to test #15250 on that. (doesn't block merge, I guess since it could be fixed before release if there were an issue)
19:12:49 <gribble> https://github.com/bitcoin/bitcoin/issues/15250 | Use RdSeed when available, and reduce RdRand load by sipa · Pull Request #15250 · bitcoin/bitcoin · GitHub
19:13:09 <wumpus> which already has some ACKs
19:13:26 <luke-jr> gmaxwell: my most recent x86 is Haswell, and it doesn't have a 32-bit chroot at the moment
19:13:46 <wumpus> gmaxwell: I consider the rdrand one a bugfix not a feature so it doesn't strictly need to be merged tomorrow
19:13:49 <gmaxwell> haswell is ducky and has both rdrand and rdseed IIRC.
19:14:09 <gmaxwell> (it certantly has rdrand)
19:14:24 <gmaxwell> wumpus: thanks, I agree.
19:14:45 <luke-jr> if our static binaries were actually portable, I could have probably tested with those, but.. :/
19:15:03 <sipa> wumpus: would #15402 count as a bugfix?
19:15:05 <gribble> https://github.com/bitcoin/bitcoin/issues/15402 | Granular invalidateblock and RewindBlockIndex by sipa · Pull Request #15402 · bitcoin/bitcoin · GitHub
19:15:16 <wumpus> sipa: I guess so
19:15:30 <wumpus> sipa: it fixes that out of memory issue on reorg right?
19:15:36 <gmaxwell> It should, it's a pretty serious bug.
19:15:39 <provoostenator> It fixes an OOM
19:15:45 <sipa> hmm, not reorg
19:16:00 <gmaxwell> wumpus: on rewind/replay
19:16:16 <wumpus> so I mean it fixes  #14289
19:16:18 <gribble> https://github.com/bitcoin/bitcoin/issues/14289 | Unbounded growth of scheduler queue · Issue #14289 · bitcoin/bitcoin · GitHub
19:16:21 <wumpus> so it's a fix
19:16:54 <wumpus> added 0.18 milestone
19:17:41 <jamesob> it doesn't fix the general issue of #14289, just one specific instance right?
19:17:42 <gribble> https://github.com/bitcoin/bitcoin/issues/14289 | Unbounded growth of scheduler queue · Issue #14289 · bitcoin/bitcoin · GitHub
19:17:50 <sipa> jamesob: indeed
19:18:25 <jamesob> intend on building in a SERIAL_SCHEDULER_EXECUTION flag (or something) for use in CI in the next couple days
19:18:29 <jamesob> (offtopic, sorry)
19:19:01 <wumpus> #topic minimum QT version for v0.18 (provoostenator)
19:19:02 <phantomcircuit> gmaxwell, that cant be checked in a vm?
19:19:20 <gmaxwell> phantomcircuit: absolutely can be
19:19:40 <luke-jr> [19:03:49] <luke-jr> I also don't see where/if anyone done a check for what versions come with current stable distros.
19:19:49 <luke-jr> I tried doing ^ myself, but I couldn't figure out what version RHEL 7 has
19:19:50 <provoostenator> I think we should either make Travis test older versions of QT or bump the min version.
19:19:54 <MarcoFalke> luke-jr: The benefit is that it is easier to just bump the minimum than to make it compile with 5.2
19:20:02 <provoostenator> Currently it doesn't even compile in 5.2
19:20:06 <MarcoFalke> ^
19:20:18 <wumpus> 5.2 is ancient anyway
19:20:22 <luke-jr> but who does this cut off?
19:20:31 <luke-jr> what version does RHEL 7 have?
19:20:34 <MarcoFalke> If someone wants to build on an acient system with an EOL qt, they are better off with depends
19:20:45 <provoostenator> As for stable distros: we can narrow that down to stable distros where the depends system doesn't work?
19:20:47 <luke-jr> RHEL 7 is current
19:20:48 <wumpus> didn't we have a discussion about this in a github issue?
19:20:57 <jonasschnelli> Yes
19:20:59 <luke-jr> provoostenator: depends should not be the norm
19:21:04 <MarcoFalke> luke-jr:  rhel7 can use depends
19:21:05 <wumpus> is there anything new to add to that or is this just another repeat of that?
19:21:13 <jonasschnelli> I think just updating Qt for the sake of updating seems pointless
19:21:15 <wumpus> this comes up too often
19:21:28 <jonasschnelli> What we need is advantages/disadvantage comparison
19:21:39 <luke-jr> updating is fine. cutting off people who reasonably have an old version isn't.
19:21:43 <meshcollider> Hi, sorry im late
19:21:43 <wumpus> jonasschnelli: he isn't updating qt, only changing the minimum requirement in the build system to the actual minimum requirement IIRC
19:22:08 <provoostenator> jonasschnelli: it's not for the sake of updating, it's because we're constantly breaking old versions.
19:22:08 <wumpus> it doesn't work with qt 5.2, no one is testing it with that, so it's kind of a moot point
19:22:15 <gmaxwell> Just pushing to new versions for the sake of updating is a bad pattern.  If there is a problem or a new feature, then great.  Matching minimums to the actual minimums is also good (excepting in cases where we should fix the minimum because we're cutting something out people are currently using)
19:22:24 <wumpus> no one is updating qt !
19:22:28 <jonasschnelli> I see. Sorry for the confusion.
19:22:29 <wumpus> this isn't the point of tha pr
19:23:08 <wumpus> #15393 bumps the minimum version checked by the bulid system
19:23:11 <gribble> https://github.com/bitcoin/bitcoin/issues/15393 | build: Bump minimum Qt version to 5.5.1 by Sjors · Pull Request #15393 · bitcoin/bitcoin · GitHub
19:23:16 <wumpus> to the version that we actually require already
19:23:23 <wumpus> the only difference is that configure will show an error now
19:23:25 <jonasschnelli> Qt 5.6 LTS ends in 2 months... :/
19:23:27 <wumpus> instead of failing later in the build
19:23:50 <jonasschnelli> Whatever minimal Qt support we set, we should make sure its ran in our CI
19:23:54 <luke-jr> can someone with a reasonable internet connection download CentOS 7 and find out what version it has? :x
19:24:00 <wumpus> jonasschnelli: yes
19:24:18 <jonasschnelli> I'm not sure if we should look to deep into all distros...
19:24:32 <MarcoFalke> luke-jr: It is 5.9?
19:24:34 <MarcoFalke> http://mirror.centos.org/centos/7/os/x86_64/Packages/qt5-qtbase-5.9.2-3.el7.x86_64.rpm
19:24:37 <jonasschnelli> If CentOS uses Qt5.3 which its support has probably been stopped 2 years ago, it's not a pattern to follow
19:24:42 <jonasschnelli> (assuming)
19:24:51 <luke-jr> MarcoFalke: not sure why I couldn't find that.. thanks
19:24:55 <sipa> luke-jr: is 5.9.2 possible? http://mirror.centos.org/centos/7/os/x86_64/Packages/
19:25:18 <jonasschnelli> make sense... 5.9 is the one where LTS is at least 1 year
19:25:28 <MarcoFalke> Is rhel7 using the same as centos7?
19:25:39 <luke-jr> MarcoFalke: CentOS is just a free-of-charge RHEL
19:25:47 <provoostenator> luke-jr: CentOS 7 seems to have QT5.9.2: http://mirror.centos.org/centos/7/os/x86_64/Packages/
19:25:53 <luke-jr> sounds like this is a non-issue then
19:25:56 <provoostenator> (oh, sipa found it)
19:25:57 <MarcoFalke> Ok, so it seems we are not cutting off anyone with the bump?
19:26:41 <wumpus> ok
19:26:48 <jonasschnelli> We cut off debian 8 which is LTS till 2020
19:26:51 <provoostenator> Cool, let's merge it right now :-P
19:26:53 <jonasschnelli> but I guess thats okay
19:27:07 <wumpus> let's merge it or close it and not talk about this for two years or so
19:27:12 <jonasschnelli> hehe
19:27:13 <jonasschnelli> indeed
19:27:26 <luke-jr> Debian 8 is oldstable?
19:28:11 <luke-jr> yeah
19:28:23 <wumpus> Debian 8 ("jessie") — obsolete stable release
19:28:26 <wumpus> obsolete
19:29:17 <MarcoFalke> Then, please someone add a debian_8 travis job. Otherwise this will be broken again
19:29:20 <jonasschnelli> yes... but we need to expect users run it close to the LTS end... but I'm fine with setting Qt 5.5.1 as min
19:29:28 <wumpus> I guess it's possible to build through depends still, but we can't support old versions of distributions that are already known to carry old packages forever
19:29:32 <luke-jr> MarcoFalke: we don't need to support oldstable IMO
19:29:41 <wumpus> no, we don't need a debian 8 travis job
19:29:44 <jonasschnelli> Yes. We have the depends build for those
19:29:53 <MarcoFalke> ok, action merge
19:30:40 <jonasschnelli> ack
19:30:43 <wumpus> ack
19:30:47 <wumpus> any other topics?
19:31:51 <bitcoin-git> [13bitcoin] 15jonasschnelli pushed 4 commits to 06master: 02https://github.com/bitcoin/bitcoin/compare/9b1df4ba0bec...758c6d784da0
19:31:52 <bitcoin-git> 13bitcoin/06master 149975282 15Luke Dashjr: GUI: If BIP70 is disabled, attempt to fall back to BIP21 parsing
19:31:52 <bitcoin-git> 13bitcoin/06master 14113f000 15Luke Dashjr: GUI: If BIP70 is disabled, give a proper error when trying to open a payme...
19:31:52 <bitcoin-git> 13bitcoin/06master 1484f5315 15Luke Dashjr: Travis: Add test without BIP70 (but still full wallet + tests)
19:32:21 <bitcoin-git> [13bitcoin] 15jonasschnelli merged pull request #15063: GUI: If BIP70 is disabled, attempt to fall back to BIP21 parsing (06master...06bip70_fallback_to_bip21) 02https://github.com/bitcoin/bitcoin/pull/15063
19:33:57 <wumpus> I guess not
19:34:14 <wumpus> #endmeeting