< ossifrage> I just built the master and it seems to be throwing the wrong error: "Fatal LevelDB error: IO error: /.../blocks/index/121486.ldb: No space left on device"
< ossifrage> however, if I stat that file, it doesn't exist
< ossifrage> and the device does not seem to be full
< wumpus> that probably means the file doesn't exist yet, it's unable to create it
< ossifrage> wumpus, I created a few test files in .../index without problems
< ossifrage> ah, found an earlier error: "BerkeleyEnvironment::Close: Error -30974 closing database environment: DB_RUNRECOVERY: Fatal error, run database recovery"
< wumpus> sounds similar to #17392
< gribble> https://github.com/bitcoin/bitcoin/issues/17392 | Fake data corruption after not really running out of disk space during IBD (pruned, Windows VM on macOS) . Issue #17392 . bitcoin/bitcoin . GitHub
< wumpus> but it definitely seems something with the disk or file system, wether a sw or hw or virtualization issue, anything weird in your kernel log?
< ossifrage> the last time I ran bitcoin-qt it ran out of disk space
< ossifrage> (sorry, that was testnet, this is mainnet, the last time my mainnet node was shutdown cleanly)
< bitcoin-git> [bitcoin] fanquake opened pull request #17916: windows: Enable heap terminate-on-corruption (master...win_set_heap_terminate_on_corruption) https://github.com/bitcoin/bitcoin/pull/17916
< wumpus> it'd only happen if the disk filled suddenly, or the OS starts returning "no space left on device" errors as false alarm; normally bitcoin would detect low disk space far before that happens and shutdown cleanly
< ossifrage> wumpus, it looks like btrfs has shat itself
< wumpus> ohh
< fanquake> I assume vcpkg has changed something boost package wise
< fanquake> Guess we need to add boost-variant to the package list. Looks like it's been split out for a while now though, so dunno why it's only just broken
< wumpus> ohnoo did vcpkg change under us again
< fanquake> ? I guess so. I assume it must also have something todo with the dependency cache, as too why it doesn't break straight away
< wumpus> right--I don't think appveyor immediately updates vcpkg in sync with upstream
< wumpus> it would be better if it was possible to fix certain versions, like we do with travis
< elichai2> fanquake: do I remember correctly that you had a PR to bump the QT version? If so I can't find it
< fanquake> elichai2: I started working on a 5.12 branch in https://github.com/fanquake/bitcoin/tree/qt_5_12_6
< elichai2> fanquake: Thanks. apparently we'll need that to compile with C++17 for some reason https://stackoverflow.com/a/49192230/1621927
< fanquake> elichai2: I'm about to post on the c++17 pr.
< elichai2> :D
< wumpus> isn't it too early for a c++17 PR?
< fanquake> wumpus: that's what I'm about to say
< elichai2> wumpus: agree
< elichai2> I thought it was a PoC
< wumpus> really hope that for 0.21 it's possible but 0.20 is first
< fanquake> Yea I think we've got plenty of other things we can worry about for at least the next 6 months
< wumpus> not that I mind anyone experimenting with C++17 already, but just to set realistic expectations with regard to merging
< bitcoin-git> [bitcoin] laanwj pushed 2 commits to master: https://github.com/bitcoin/bitcoin/compare/e258ce792a48...f2f9fdf5794e
< bitcoin-git> bitcoin/master 831e122 fanquake: build: remove double LIBBITCOIN_SERVER linking
< bitcoin-git> bitcoin/master f2f9fdf Wladimir J. van der Laan: Merge #17910: build: remove double LIBBITCOIN_SERVER linking
< bitcoin-git> [bitcoin] laanwj merged pull request #17910: build: remove double LIBBITCOIN_SERVER linking (master...remove_double_server) https://github.com/bitcoin/bitcoin/pull/17910
< bitcoin-git> [bitcoin] laanwj pushed 3 commits to master: https://github.com/bitcoin/bitcoin/compare/f2f9fdf5794e...2ed74a43a05a
< bitcoin-git> bitcoin/master b17e91d James O'Beirne: refactoring: introduce CChainState::GetCoinsCacheSizeState
< bitcoin-git> bitcoin/master 02b9511 James O'Beirne: tests: add tests for GetCoinsCacheSizeState
< bitcoin-git> bitcoin/master 2ed74a4 Wladimir J. van der Laan: Merge #16945: refactor: introduce CChainState::GetCoinsCacheSizeState
< bitcoin-git> [bitcoin] laanwj merged pull request #16945: refactor: introduce CChainState::GetCoinsCacheSizeState (master...2019-09-au-shouldflush) https://github.com/bitcoin/bitcoin/pull/16945
< sipsorcery> re. appveyor CI the vs2019 image was updated on 9th of Jan.
< sipsorcery> I checked the build against my fork and all was and still is good.
< fanquake> sipsorcery: what do you think might be the issue in https://ci.appveyor.com/project/DrahtBot/bitcoin/builds/30069264 ?
< sipsorcery> fanquake: first guess would be CI gremlins. The fact that the job failed within 8 mins is very suspicious. Way too quick.
< sipsorcery> There's been a successful build since then so it'd be worth re-triggering that PR.
< fanquake> I'll do that now
< bitcoin-git> [bitcoin] practicalswift closed pull request #17904: [wip] build: Switch from C++11 to C++17 (master...c++17) https://github.com/bitcoin/bitcoin/pull/17904
< luke-jr> hebasto: ryanofsky: this is the current Knots intro (before Knots branding) http://luke.dashjr.org/tmp/screenshots/Screenshot_20200113_081940.png
< bitcoin-git> [bitcoin] practicalswift opened pull request #17917: tests: Add amount compression/decompression fuzzing to existing fuzzing harness (master...amount-compression-roundtrip) https://github.com/bitcoin/bitcoin/pull/17917
< hebasto> luke-jr: ty
< luke-jr> hebasto: ryanofsky: FWIW, current code is at rwconf_gui-0.19+knots if you're looking to rebase
< hebasto> sipsorcery: do you mean AppVeyor vs2019 image with pre-installed Qt? https://www.appveyor.com/blog/2019/10/11/vs2019-image-with-docker-wsl-preheated/
< sipsorcery> hebasto: the appveyor image has always had Qt pre-installed but it's not statically linked so can't be used by us.
< hebasto> sipsorcery: ty, I got it already/
< luke-jr> sipsorcery: there's no reason we should require a static Qt..
< luke-jr> especially for CI
< sipsorcery> luke-jr: true but the original intent was to replicate he existing build processes and they are all static.
< luke-jr> they're also all non-MSVC ;)
< bitcoin-git> [bitcoin] emilengler opened pull request #17918: qt: Hide non PKHash-Addresses in signing address book (master...2020-01-hide-non-pkhash-addresses) https://github.com/bitcoin/bitcoin/pull/17918
< luke-jr> FWIW, to avoid all the /merge branches, you can configure the origin-pull remote as: fetch = +refs/pull/*/head:refs/remotes/origin-pull/*/head
< bitcoin-git> [bitcoin] dongcarl opened pull request #17919: depends: Allow building with system clang (master...2020-01-depends-macos-allow-system-clang) https://github.com/bitcoin/bitcoin/pull/17919
< bitcoin-git> [bitcoin] dongcarl opened pull request #17920: guix: Build support for macOS (master...2020-01-guix-macos) https://github.com/bitcoin/bitcoin/pull/17920
< bitcoin-git> [bitcoin] theStack opened pull request #17921: test: test OP_CSV empty stack fail in feature_csv_activation.py (master...20200113-test-check-for-empty-stack-op_csv) https://github.com/bitcoin/bitcoin/pull/17921