=== michihenning is now known as michi
=== michi is now known as Guest23427
[11:18] <Laney> hmm
[11:18] <Laney> Is it expected that things being built against the s390x bootstrap archive are getting held in proposed?
[11:19] <Laney> For example texlive-binaries got a dependency on libpoppler56 which it got from the bootstrap archive, and now britney is (correctly) saying it is uninstallable
[11:21] <Laney> ...It would actually want a rebuild for the poppler transition on s390x, but poppler isn't built yet
[11:58] <cjwatson> Mirv: Looks like qtbase-opensource-src build-depends on publicsuffix, which is in universe.  Could you sort out an MIR?
[11:58] <cjwatson> That's needed for Laney's request, via a couple of layers
[11:58] <Laney> I just asked him in #ubuntu-desktop
[11:58] <Laney> but thanks
[11:58] <Laney> Think it should be a trivial MIR
[11:59] <Mirv> cjwatson: darn, yes, I'll file a MIR
[11:59] <infinity> I could toss a bootstrap build of qtbase-opensource-src from the release pocket into a bootstrap repo.  Maybe.
[11:59] <Laney> You could pre-promote publicsuffix based on its triviality ;)
[12:00] <xnox> Mirv, cause without that, i can't build qtbase-opensource-src in proposed on s390x, and then i can't build new poppler, and the world is sad, and -proposed will pile up =)
[12:00] <infinity> Laney: Looking now.
[12:00] <Laney> Ta
[12:01] <infinity> Yeah.
[12:01] <infinity> I'm not on the team anymore, but happy to pre-promote.
[12:01] <infinity> It's a single data file. :P
[12:01] <cjwatson> And sigh, if I'd realised ghc and haskell-* were going to be uploaded overnight I'd have sync-blacklisted them.
[12:01] <cjwatson> Now I get to rebuild everything.
[12:01] <xnox> infinity, NB! _and_ a symlink =)
[12:03] <Laney> Hopefully qtbase builds
[12:03] <infinity> It had a single test failure in the previous version.
[12:03] <infinity> Maybe this one's happier.
[12:03] <infinity> Fingers crossed.
[12:04] <xnox> yeah, there were weird QtFile test suite failures, with long 240+ characters file names, on a debian kernel...
[12:04] <xnox> and different toolchain...
[12:05] <xnox> and i see more qtbase being built in landing-032 again ?!
[12:07] <Laney> Silos might need s390x turned on, if that's an explicit knob
[12:08] <xnox> wgrant, ^
[12:08] <xnox> Laney, i think it is a knob
[12:09] <Mirv> Laney: xnox: cjwatson: https://bugs.launchpad.net/ubuntu/+source/publicsuffix/+bug/1522811
[12:10] <ubot5> Ubuntu bug 1522811 in publicsuffix (Ubuntu) "[MIR] publicsuffix" [Undecided,New]
[12:10]  * Laney adds s390x to transition-tracker
[12:10] <Laney> Mirv: ta
[12:14] <cjwatson> Laney: It is, but we want to let the bootstrap finish first
[12:14] <cjwatson> There's always going to be a slightly ambiguous period
[12:15] <cjwatson> But let's not confuse people with a bazillion dep-waits
[12:18] <Laney> As you wish
[12:22] <infinity> I was probably going to turn on s390x in silos early next week.
[12:22] <infinity> Once we've made sure most of our stack kinda builds.
[12:29] <xnox> building qtbase-opensource-src in a nonvirt ppa, with universe enabled, s390x build to see if it will build. (thus not waiting publicsuffix mir)
[12:55] <infinity> xnox: it's already building in the primary archive.
[12:57] <Laney>  Totals: 55 passed, 1 failed, 0 skipped, 0 blacklisted
[12:57] <Laney> ********* Finished testing of tst_LargeFile *********
[12:57] <Laney> Makefile:573: recipe for target 'check' failed
[13:01] <xnox> as usual.
[13:01] <xnox> oh but qfile did pass.
[13:02] <infinity> #if defined(__x86_64__)
[13:02] <infinity>         QEXPECT_FAIL("", "fails on 64-bit Linux (QTBUG-21175)", Abort);
[13:02] <infinity> #endif
[13:02] <infinity> WORST IFDEF EVER.
[13:02] <infinity> Fails on 64-bit Linux, so we'll exclude only one 64-bit arch.
[13:04] <infinity> And before anyone asks "but, how does it work on arm64 and ppc64el then?"
[13:04] <infinity> # Skip tests on the archs they are known to be flaky  with current configuration
[13:04] <infinity> testskip_architectures := arm64 powerpc ppc64el
[13:04] <infinity> Brilliant in so many ways.
[13:05] <infinity> Anyhow, someone fix either the ifdef in the test (preferred) or the excluded test arches (ick), and you're golden.
[13:05] <infinity> Mirv, xnox: ^
[13:05] <xnox> infinity, i have chroot up with qtbase, i was working on excluding individual tests, and marking htem expect fail
[13:05] <xnox> infinity, cause most tests do pass.
[13:06] <infinity> xnox: There's only one failing test.
[13:06] <infinity> xnox: And there's already an ifdef there to exclude it on amd64.  That's what I was pointing out. :P
[13:06] <xnox> infinity, thanks. i was fixing broken schroot on debian, that doesn't support too-new kernel with too new overlayfs expected options =)
[13:07] <infinity> xnox: Use aufs?
[13:07] <xnox> infinity, and there might be more tests after it, as the build is done without -k
[13:07] <xnox> infinity, phhhh, i'm hipster.
[13:09] <Mirv> infinity: xnox: my next upload will make the archs to run tests on opt-in instead of opt-out, so xnox can then later eg add s390x when it passes
[13:09] <xnox> Mirv, please no, i want to run tests on s390x.
[13:10] <xnox> or keep it enabled, i'll upload direct to archive, rather than silo to get it building and migrating.
[13:10] <xnox> most of them pass on s390x, it's not a new platform, but a solid one.
[13:10] <cjwatson> Well, opt-in is still good, you can just add s390x to the list.
[13:10] <Mirv> xnox: well it's better to specify which architectures to run on, but I'll gladly add s390x there if you have the changes by that time
[13:10] <xnox> Mirv, yeap.
[13:12] <infinity> Mirv: The line I pointed out above is obviously wrong anyway.
[13:13] <infinity> Mirv: When the comment says "doesn't work on 64-bit" and then it only excludes on 64-bit arch. :P
[13:13] <infinity> Mirv: Especially since Qt pretty much has to have some is_64bit macro or something that's used in, like, the entire codebase...
[13:13] <infinity> s/excludes on/excludes one/
[13:19] <Mirv> infinity: sure it's wrong. upstream CI is x86+armhf only so they probably do things like that in tests at times unfortunately
[13:33] <cjwatson> right, I'm going to cancel all the haskell builds that are still "needs building", partly to get queues under control a bit more and partly because they're mostly going to need rebuilds anyway
[15:02] <xnox> proposed-migrations says that zipl-installer is not installable, yet i don't see how or why =(
[15:14] <cjwatson> xnox: with all the virtual dependencies in use in the installer, it's probably fallout from something earlier on, maybe partman-base not being installable
[15:16] <cjwatson> xnox: a lot of things should get less ridiculously difficult to unpick once a new kernel lands
[15:17] <xnox> ack.
[15:17] <xnox> well i added -proposed and -release, debian-installer bits and e.g. zipl-installer is installable, so it's okish. i guess it just doesn't migrate on it's own then.
[15:19] <cjwatson> xnox: It transitively depends on kernel udebs, so until they're migratable, migrating zipl-installer would increase the installability count by one
[15:20] <cjwatson> Even though it's basically for an existing reason
[16:21] <apw> cjwatson, fyi i just released the 4.3 kernel
[16:24] <cjwatson> apw: oh good, thanks
[16:50] <xnox> where is autopackage test page?
[16:50] <xnox> ...
[16:50] <xnox> why did i cause new binaries for procps.... /me looks
[16:51] <xnox> abi bump, sigh.
[17:47] <xnox> E: Invalid Release signature (key id 40976EAF437D05B5) -> i hate the world
[20:07] <xnox> tst_QImage 1 failed, tst_QtJson 1 failed, tst_QLogging 11 failed,
[20:07] <xnox> =(