--- Log opened Tue Nov 05 00:00:58 2019
07:49 -!- jb55 [~jb55@gateway/tor-sasl/jb55] has quit [Ping timeout: 260 seconds]
07:52 -!- jb55 [~jb55@gateway/tor-sasl/jb55] has joined #bitcoin-builds
08:39 < dongcarl> wumpus: MarcoFalke: did we recently bump univalue?
08:39 < fanquake> dongcarl: https://github.com/bitcoin/bitcoin/pull/17324
08:40 < dongcarl> Ah but MarcoFalke's bisecting shows that the subtree update didn't cause this...
08:40 < dongcarl> fanquake: We're talking about https://github.com/bitcoin/bitcoin/issues/17323 btw
08:41 < fanquake> dongcarl: I've been following somewhat. Seems weird that this has just started happening (or no one noticed until) now.
08:42 < dongcarl> fanquake: True. I've been Gitian-building a lot of the build stuff
08:46 < dongcarl> I'm building v0.19.0rc3 over and over right now, just to see if I can reproduce
08:54 < dongcarl> MarcoFalke: Is the failure still sporadic?
10:42 < wumpus> I think the --recheck is a good observation dongcarl , this is something that might be triggered by something date/time related
10:42 < wumpus> and it's true, a recheck in a subdir won't pick up the site config script
10:43 < dongcarl> Afk right now but Cory says we should look at his univalue c++11 PR
10:43 < wumpus> where?
10:44 < dongcarl> I'll try to find it he's shown us before
10:46 < dongcarl> wumpus: https://github.com/bitcoin/bitcoin/pull/12467
10:47 < wumpus> I don't think that has anything to do with this particular problem
10:47 < wumpus> that'd definitely not cause intermittent issues
10:48 < wumpus> I considered it at first whether it'd be a ABI conflict but it wouldn't make sense
10:49 < dongcarl> Looking at the first and third commits in this PR, it seems relevant. But I can't figure out why it'd be intermittent either.
10:49 < dongcarl> Might be the wrong clue
10:51 < MarcoFalke> I hope that my bisect is working 0.18.1 succeeded 20 times, then I aborted. I could be out of luck and all builds fail ...
10:51 < MarcoFalke> ... eventually
10:52 < MarcoFalke> Currently running a loop on https://github.com/bitcoin/bitcoin/commit/1a8a5ede9f, which is in both 0.18 branch and 0.19 branch
11:12 < MarcoFalke> ok, that failed :(
11:52 < wumpus> so, we should either make sure an automatically triggered recheck on univalue works, or ensure it never happens
11:53 < wumpus> the former is probably easier
11:53 < wumpus> I can think of a zillion ways to trigger the latter
11:54 < dongcarl> true
11:54 < dongcarl> What normally automatically triggers rechecks?
11:55 < dongcarl> automatically = "smart"ly = intermittently
12:32 < MarcoFalke> 0.18.1 also fails with the 0.18.1 descriptor (secp)
12:33 < MarcoFalke> Will upload a dump shortly
12:33 < dongcarl> MarcoFalke: Could you include full build.log with the dump?
12:33 < MarcoFalke> Sure
12:34 < dongcarl> My gut feeling is that our FAKETIME wrapping is messing with Make's assumptions somehow...
12:36 -!- jonatack [~jon@2a01:e35:8aba:8220:6627:dad:d967:649d] has quit [Quit: jonatack]
12:45 < MarcoFalke> Given that pretty much all commits of Bitcoin Core fail to build, but I've never seen this in the past, this is probably an issue with podman/docker
12:46 < dongcarl> Or perhaps the Bionic image?
12:49 < MarcoFalke> jup
12:49 < MarcoFalke> Though, a bit strange that no one else is seeing these
12:51  * dongcarl going to the ubuntu bug tracker
12:53 < fanquake> There was a new release of faketime fairly recently
12:57 < dongcarl> Bionnic is still on the old release I think https://packages.ubuntu.com/bionic/libfaketime
12:57 < fanquake> yep, no changes for 1.5 years
12:57 < wumpus> dongcarl: changes in univalue's configure.ac at least
12:58 < wumpus> (maybe changes in configure, or other build system files)
12:58 < fanquake> This was the new release: https://github.com/wolfcw/libfaketime/releases/tag/v0.9.8
12:59 < dongcarl> I believe I know _how_ these failures are happening, but not _why_... I'll upload my thoughts on the thread in case it's useful
12:59 < MarcoFalke> Is there inferiours for bionic apt-get? :eyes:
13:00 < MarcoFalke> *Inferiors
13:01 -!- jonatack [~jon@2a01:e35:8aba:8220:6627:dad:d967:649d] has joined #bitcoin-builds
13:01 < dongcarl> MarcoFalke: I think so? But you gotta know the version number
13:01 < dongcarl> And the server's gotta still have it lying around
13:05 < MarcoFalke> reminds me that the gitian builds should be bumped to Ubuntu Focal 20.04
13:06 < MarcoFalke> will do that now
13:06 < fanquake> Is that even released
13:06 < fanquake> or am I misunderstanding what you mean
13:06 < dongcarl> I think Marco is being Marco
13:09 < dongcarl> I promise I'm trying my best to get OSX Guix builds working... Hopefully we don't have to bump to focal and write all our glibc compat hooks again......
13:12 < MarcoFalke> I'd say we should keep gitian for 0.20.0
13:14 < MarcoFalke> It's a scripted-diff: https://github.com/MarcoFalke/bitcoin-core/commits/1911-buildFocal
13:14  * MarcoFalke hides
13:15 < MarcoFalke> Let's see if the build failures are also happening with that commit
13:15 < dongcarl> MarcoFalke: That definitely won't work unfortunately
13:16 < MarcoFalke> too bad
13:16 < MarcoFalke> fanquake: Not released, but the image is available
15:54 -!- jonatack [~jon@2a01:e35:8aba:8220:6627:dad:d967:649d] has quit [Ping timeout: 245 seconds]
16:31 -!- alko89 [~alko@cpe-85-10-28-138.static.amis.net] has quit [Ping timeout: 268 seconds]
16:54 -!- jb55 [~jb55@gateway/tor-sasl/jb55] has quit [Quit: jb55]
17:09 -!- jb55 [~jb55@gateway/tor-sasl/jb55] has joined #bitcoin-builds
22:55 -!- alko89 [~alko@cpe-85-10-28-138.static.amis.net] has joined #bitcoin-builds
--- Log closed Wed Nov 06 00:00:58 2019