[01:53] -queuebot:#ubuntu-release- Unapproved: xapers (disco-proposed/universe) [0.8.2-1 => 0.8.2-1ubuntu1] (no packageset)
[02:02] -queuebot:#ubuntu-release- Unapproved: aioprocessing (disco-proposed/universe) [1.0.1-1 => 1.0.1-1ubuntu1] (no packageset)
[02:15] -queuebot:#ubuntu-release- Unapproved: botch (disco-proposed/universe) [0.21-6 => 0.21-6ubuntu1] (no packageset)
[02:54] -queuebot:#ubuntu-release- Unapproved: pep8 (disco-proposed/main) [1.7.1-1ubuntu1 => 1.7.1-1ubuntu2] (kubuntu, ubuntu-desktop)
[02:59] -queuebot:#ubuntu-release- Unapproved: pygresql (disco-proposed/universe) [1:5.0.6-1build1 => 1:5.0.6-1ubuntu1] (no packageset)
[03:39] -queuebot:#ubuntu-release- Unapproved: iproute2 (xenial-proposed/main) [4.3.0-1ubuntu3.16.04.3 => 4.3.0-1ubuntu3.16.04.4] (core)
[06:55] -queuebot:#ubuntu-release- Unapproved: gcc-6-cross (disco-proposed/universe) [30ubuntu4 => 33ubuntu1] (no packageset)
[06:59] -queuebot:#ubuntu-release- Unapproved: gcc-6-cross-ports (disco-proposed/universe) [28ubuntu5 => 30ubuntu1] (no packageset)
[07:02] -queuebot:#ubuntu-release- Unapproved: accepted aioprocessing [source] (disco-proposed) [1.0.1-1ubuntu1]
[07:02] -queuebot:#ubuntu-release- Unapproved: accepted gcc-6-cross-ports [source] (disco-proposed) [30ubuntu1]
[07:02] -queuebot:#ubuntu-release- Unapproved: accepted pep8 [source] (disco-proposed) [1.7.1-1ubuntu2]
[07:02] -queuebot:#ubuntu-release- Unapproved: accepted xapers [source] (disco-proposed) [0.8.2-1ubuntu1]
[07:02] -queuebot:#ubuntu-release- Unapproved: accepted botch [source] (disco-proposed) [0.21-6ubuntu1]
[07:02] -queuebot:#ubuntu-release- Unapproved: accepted pygresql [source] (disco-proposed) [1:5.0.6-1ubuntu1]
[07:02] -queuebot:#ubuntu-release- Unapproved: accepted gcc-6-cross [source] (disco-proposed) [33ubuntu1]
[07:27] -queuebot:#ubuntu-release- Unapproved: valgrind (disco-proposed/main) [1:3.14.0-0ubuntu1 => 1:3.14.0-0ubuntu2] (ubuntu-desktop)
[07:29] -queuebot:#ubuntu-release- Unapproved: accepted valgrind [source] (disco-proposed) [1:3.14.0-0ubuntu2]
[07:42] -queuebot:#ubuntu-release- Unapproved: botch (disco-proposed/universe) [0.21-6ubuntu1 => 0.21-6ubuntu2] (no packageset)
[08:02] -queuebot:#ubuntu-release- Unapproved: accepted botch [source] (disco-proposed) [0.21-6ubuntu2]
[08:02] <mwhudson> coreycb: oh hey do you want to update murano for python 3.7? :)
[08:03] <mwhudson> oh wait it's fixed upstream
[08:12] -queuebot:#ubuntu-release- Unapproved: libhtml-html5-sanity-perl (disco-proposed/universe) [0.105-3 => 0.105-4] (no packageset) (sync)
[08:17] -queuebot:#ubuntu-release- Unapproved: accepted libhtml-html5-sanity-perl [sync] (disco-proposed) [0.105-4]
[08:32] -queuebot:#ubuntu-release- Unapproved: python-vobject (disco-proposed/universe) [0.9.5-2 => 0.9.6.1-0.1] (kubuntu) (sync)
[08:34] -queuebot:#ubuntu-release- Unapproved: hovercraft (disco-proposed/universe) [2.1-3 => 2.1-3ubuntu1] (no packageset)
[08:38] -queuebot:#ubuntu-release- Unapproved: accepted hovercraft [source] (disco-proposed) [2.1-3ubuntu1]
[08:38] -queuebot:#ubuntu-release- Unapproved: accepted python-vobject [sync] (disco-proposed) [0.9.6.1-0.1]
[08:45] -queuebot:#ubuntu-release- Unapproved: valgrind (disco-proposed/main) [1:3.14.0-0ubuntu2 => 1:3.14.0-0ubuntu3] (ubuntu-desktop)
[08:50] -queuebot:#ubuntu-release- Unapproved: accepted valgrind [source] (disco-proposed) [1:3.14.0-0ubuntu3]
[09:03] <Laney> doko: dunno, is the problem fixed?
[09:05] <LocutusOfBorg> doko, good morning, llvm-defaults can migrate if you remove (NBS in proposed cleanup) libc++-test, libiomp-dev, libc++abi-test, all removed in debian until the properly start working
[09:11] <doko> LocutusOfBorg: these do not exist in -proposed
[09:21] <xnox> jbicha, well, before zipl, you can see that upgrading perl-base removes perl, ubuntu-minimal, s390-tools et.al. which is not good.
[09:23] <LocutusOfBorg> doko, NBS in proposed, so you have to clean them up in release?
[09:39] <doko> 0.43ubuntu1 doesn't have these either
[09:49] <mwhudson> wtf https://launchpadlibrarian.net/396099495/buildlog_ubuntu-disco-s390x.botch_0.21-6ubuntu3~ppa1_BUILDING.txt.gz
[09:51] <xnox> mwhudson, i've seen that before
[09:52] <xnox> also i'm not sure if botch tests are just a pure botch or not.
[09:52] <mwhudson> ah out/acl-ma-path.dot is empty in my non-purged local sbuild failure
[09:54] <mwhudson> the command that was supposed to produce it certainly produced output when i ran it again
[10:34] <LocutusOfBorg> doko, previous llvm-toolchain-6.0 had them
[10:36] -queuebot:#ubuntu-release- Unapproved: sssd (xenial-proposed/main) [1.13.4-1ubuntu1.11 => 1.13.4-1ubuntu1.12] (kubuntu, ubuntu-desktop, ubuntu-server)
[10:41] <doko> the unversioned packages?
[10:43] <Laney> LocutusOfBorg: what's with those armhf requests?
[10:43] <doko> trying: llvm-defaults
[10:43] <doko> skipped: llvm-defaults (0, 2, 7)
[10:43] <doko>     got: 24+0: a-2:a-2:a-13:i-2:p-2:s-3
[10:43] <doko>     * armhf: libc++-dev, libc++-test, libc++1, libc++abi-dev, libc++abi-test, libc++abi1, libiomp-dev, libiomp5-dbg, libomp-dev, libomp5, libomp5-dbg
[11:03] <LocutusOfBorg> Laney, context please?
[11:04] <LocutusOfBorg> doko, I guess so, not sure
[11:04] <doko> that's llvm-defaults not migrating
[11:06] <Laney> there were a ton of armhf requests for passed things
[11:06] <Laney> I deleted them all
[11:06] <LocutusOfBorg> Laney, I did only this: ./retry-autopkgtest-regressions --state REGRESSION | grep perl | vipe | xargs -rn1 -P10 wget --load-cookies ~/.cache/autopkgtest.cookie -O-
[11:06] <LocutusOfBorg> this is the only command I ran in the past few days
[11:06] <Laney> well you ended up requesting things that were already green
[11:06] <LocutusOfBorg> (also something for gdbm, but nvm)
[11:07] <LocutusOfBorg> Laney, how REGRESSION can pick up something that is green?
[11:07] <Laney> perhaps they passed in the previous run or you ran it twice or something
[11:07] <LocutusOfBorg> no twice, because it took an hour to run the command...
[11:08] <LocutusOfBorg> but yeah, maybe the passed between the start of the command and the end of it :)
[11:10] <LocutusOfBorg> doko, got it
[11:10] <LocutusOfBorg> https://packages.qa.debian.org/libc/libc++.html
[11:10] <LocutusOfBorg> this one has to GO AWAY WITH FIREEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE
[11:10] <LocutusOfBorg> already removed from debian, now part of llvm IIRC
[11:11] <LocutusOfBorg> look at https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=912162 to know what to remove
[11:11] <LocutusOfBorg> hint: src:libc++ libc++-helpers libc++-test libc++abi-test
[11:26] <doko> removed
[11:26] <LocutusOfBorg> ta
[11:47] -queuebot:#ubuntu-release- Unapproved: chromium-browser (disco-proposed/universe) [70.0.3538.67-0ubuntu0.18.10.1 => 70.0.3538.77-0ubuntu1] (ubuntu-desktop)
[11:56] <doko> Laney, xnox: were the s390x unknown give-backs "killed" as well?
[11:57] <Laney> yes, some of those were duplicates and i wanted to see if non all-proposed works
[11:58] <Laney> once the pending requests are finished we can see and then retry in the best way
[12:08] <doko> 144 to go
[12:11] <Laney> exciting
[12:17] <xnox> Laney, non all-proposed, triggered by perl, do not work.
[12:17] <xnox> Laney, but i am happy to wait.
[12:17] <Laney> false
[12:17] <xnox> horum
[12:17] <Laney> perl package names are hard to type
[12:18] <xnox> Laney, not triggered by perl-* but triggered by ^perl$
[12:18] <xnox> Laney, i'm yet to see any pass on s390x for http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#perl
[12:18] <xnox> (note the #perl)
[12:20] <Laney> okey dokey, let's see, doko has some of those in the queue
[12:20] <doko> yep, the first few
[12:21] <Laney> there's this thing where it's supposed to just fall back to doing that in the case of uninstallable packages
[12:21] <Laney> but I guess if apt decides to remove stuff instead of failing that won't work
[12:21] <Laney> do wonder why this would be s390x specific
[12:21] <doko> libreoffice still triggered by gcc-N ...
[12:47] <xnox> Laney, well, it seems to try to upgrade perl-base very early on. and it doesn't seem to upgrade perl or install new perl modules, and s390-tools which is installed on the cloud images depends on perl.
[12:48] <xnox> Laney, as if it chooses not to install new packages, and remove packages instead.
[12:48] <xnox> i don't think on other arches bootloader package depends on perl
[12:49] <xnox> doko, Laney - if they didn't work this morning, i don't expect them to start working this afternoon.
[12:49] <xnox> e.g. the abi-compliance-checker {"requester": "doko", "triggers": ["perl/5.28.0-3"]}
[12:58] <coreycb> mwhudson: yes will do
[13:26] <Laney> xnox: yeah OK I believe you now :P
[13:30] <Laney> i've inserted them again
[13:30] -queuebot:#ubuntu-release- Unapproved: tmux (bionic-proposed/main) [2.6-3 => 2.6-3ubuntu0.1] (ubuntu-server)
[13:44] -queuebot:#ubuntu-release- Unapproved: libreoffice (disco-proposed/main) [1:6.1.2-0ubuntu4 => 1:6.1.3-0ubuntu1] (kubuntu, ubuntu-desktop)
[13:50] -queuebot:#ubuntu-release- Unapproved: libreoffice-l10n (disco-proposed/main) [1:6.1.2-0ubuntu1 => 1:6.1.3-0ubuntu1] (ubuntu-desktop)
[13:50] <xnox> Laney, how did you insert them, yet be in the huge queue?
[13:51] <xnox> Laney, when i do `retry` script from ubuntu-archive-tools they go into `ubuntu` queue, rather than `huge`.
[13:51] <xnox> or is it because you have ssh powers?
[13:51] -queuebot:#ubuntu-release- Unapproved: python-openstackclient (bionic-proposed/main) [3.14.0-0ubuntu1 => 3.14.2-0ubuntu1] (openstack, ubuntu-server)
[15:06] <Laney> xnox: special powers
[15:07] <xnox> you are such a wonder woman
[15:09] <Laney> 🧘
[15:24] <LocutusOfBorg> I think llvm-defaults needs a special hint to migrate, can any AA please verify/do it?
[15:24] <LocutusOfBorg>     * armhf: libc++-dev, libc++1, libc++abi-dev, libc++abi1, libiomp-dev, libiomp5-dbg, libomp-dev, libomp5, libomp5-dbg
[15:25] <LocutusOfBorg> they used to be part of src:libc++ and now part of llvm-defaults
[15:25] <LocutusOfBorg>  libc++-dev | 6.0.1-1          | cosmic/universe         | amd64, arm64, armhf, i386, ppc64el, s390x
[15:25] <LocutusOfBorg>  libc++-dev | 1:7.0-44         | disco-proposed/universe | amd64, arm64, armhf, i386, ppc64el, s390x
[15:25] <doko> libc++ sources are removed in disco
[15:26] <doko> or are there more sources?
[15:26] <LocutusOfBorg> oh... let me check
[15:26] <LocutusOfBorg> indeed
[15:26] <LocutusOfBorg> cosmic!=disco
[15:26] <doko> openmprtl
[15:27] <LocutusOfBorg> indeed, that one!
[15:28] <infinity> No hint is needed for binary takeovers, nor is the removal that happened.
[15:29] <doko> intel-mkl needs syncing first
[15:34] <bdmurray> I'm having an issue with debmirror on Ubuntu 18.04. https://pastebin.canonical.com/p/sFjzWx9nC8/
[15:34] <infinity> bdmurray: pastebin.canonical in a public channel?  Naughty.
[15:34] <bdmurray> https://pastebin.ubuntu.com/p/GDNnPNcScS/
[15:35] <bdmurray> Its early man
[15:35] <LocutusOfBorg> infinity, sometimes when real binaries are superseeded by provided binaries in another src package, britney gets confused, at least I recall having to hint some llvm stuff in the past
[15:35] <LocutusOfBorg> but not in this case, indeed
[15:35] <infinity> bdmurray: That'll fix itself when xnox SRUs ubuntu-keyring.  But also, yay.
[15:35] <infinity> xnox: Brian found another thing that breaks with dual signing. :P
[15:36] -queuebot:#ubuntu-release- Unapproved: network-manager (disco-proposed/main) [1.12.4-1ubuntu1 => 1.12.4-1ubuntu1.1] (kubuntu, ubuntu-desktop) (sync)
[15:36] -queuebot:#ubuntu-release- Unapproved: systemd (disco-proposed/main) [239-7ubuntu10 => 239-7ubuntu10.1] (core) (sync)
[15:36] <bdmurray> infinity: Is there anything I can do in the meantime?
[15:36] <infinity> bdmurray: Manually add the 2018 key to your keyring on that machine.
[15:37] <infinity> bdmurray: Or fix debmirror to use less braindead signature validation (preferred).
[15:37] <infinity> bdmurray: Since "1 sig good, other sig unknown" should be "good", not "bad".
[15:37] <infinity> Andjust trusting raw gpgv returns doesn't get you the desired result.
[15:39] <infinity> xnox: Maybe what gpgv really wants here is a new commandline option to change the behaviour to the above.
[15:39] <infinity> xnox: Defaulting to behaving as it always has, but then things like debmirror could use the new option to "--ignore-unknown-if-good-sig-found" or something entirely less verbose and ugly. :P
[15:54] -queuebot:#ubuntu-release- Unapproved: curl (disco-proposed/main) [7.61.0-1ubuntu2 => 7.61.0-1ubuntu2.2] (core) (sync)
[15:58] <cjwatson> debmirror is often run on oldish machines, so I'd prefer something that works with existing gpgv
[15:59] <infinity> cjwatson: I mean, this isn't the first time we've had a dual-signed archive. :P
[15:59] <infinity> cjwatson: The solution on oldish machines with oldish gpgv is to make sure the new key is SRUed back.
[15:59] <infinity> Which, I think, xnox has planned to do ASAP.
[15:59] <infinity> But it's still crap when any tool fails in a "1 good, 1 unknown" scenario, and we should fix that going forward, IMO.
[16:00] <infinity> For smoother and less poop key rollovers.
[16:00] <infinity> I can see why gpgv's default is what it is, as for document signing, you may well want to treat any unknown as a failure, since your goal is to see if ALL the parties signed.
[16:01] <infinity> But for an apt archive, we have a pretty clearly defined thing we want, which is "if there a good sig with a key I trust, I trust it all."
[16:02] <cjwatson> I agree we should SRU back the new key - just saying that a debmirror patch that uses a shiny new gpgv option probably won't fly
[16:03] <cjwatson> At least not until it's existed for a few years and is in various stable distros
[16:06] <infinity> Oh, as in your concern is people backporting debmirror?
[16:06] <infinity> The patch could check for the option before using it, in that case.
[16:07] <xnox> infinity, bonus points: 2 good = 0; 1 good + 1 bad/expired/invalid sig = 1; 1 good + 1 nopubkey = 2. My expectation from manpage was to get return code 1, when at least one sig was god. but alas, no.
[16:08] <xnox> infinity, cjwatson - the generic solution here is to parse the status-fd / logfile of gpgv, which seems to go back in time quite well with fixed output in C.UTF-8 and is what apt-helper gpgv method is using.....
[16:08]  * xnox really wants to expose apt-helper gpgv method, as a gpgv interface. (see chatter on #ubuntu-devel)
[16:09] <infinity> xnox: Yeah, the manpage's return codes thing seems a bit odd too.  But I'd prefer explicitly asking for somehting, not trusting magic returns.
[16:09] <infinity> If what you want is "is any sig good", you should be able to ask for that.
[16:09] <xnox> infinity, that needs parsing log output as done by apt method gpgv, unfortunately.
[16:09] <xnox> infinity, i think one should be able to trivially ask for that from gpgv, which i will file separately.
[16:10] <xnox> (is any good)
[16:10] <infinity> xnox: Sure, today that's true.  My point is I should be able to ask for "if one or more sigs are good, please return 0" and get that.
[16:10] <xnox> +1
[16:10] <xnox> but also it now seems like i will need to sru ubuntu-keyring all the way back to trusty, if not esm precise too.
[16:10] <infinity> xnox: I think our error in parsing the RETURN CODES section is that "unknown" is fatal, not bad.
[16:11] <infinity> xnox: If there were 1 good and 1 bad, it would be returning 1.  I think.
[16:11] <xnox> which imho, is a side-effect that NOPUB is not a subset of signature errors like SIGBAD, SIGEXPIRED, etc.
[16:11] <xnox> inside gpg code
[16:12] <xnox> anyway, bugs everywhere, mondays are awesome
[16:12] <infinity> xnox: Yup.
[16:14] <cjwatson> infinity: I've never had good experiences with "check for option before using it" for utility programs.  Parsing the status-fd as xnox suggests seems more robust.
[16:14] <infinity> cjwatson: Yeah, don't disagree on that point.
[16:16] <LocutusOfBorg> please accept magics++, this seems to be a possible fix for the python3.7 link failure on metview (holding gdbm transition)
[16:16] -queuebot:#ubuntu-release- Unapproved: magics++ (disco-proposed/universe) [3.1.0-2build1 => 3.2.1-1] (no packageset) (sync)
[16:16] <infinity> LocutusOfBorg: Oh, I was just investigating that.  Lemme look.
[16:17] <LocutusOfBorg> infinity, seems also that magics++ needs a python3-all-dev dependency, because of this:
[16:17] <LocutusOfBorg> --    MAGICS_LIBRARIES : [MagPlus eccodes eccodes_f90 /usr/lib/x86_64-linux-gnu/libpng.so /usr/lib/x86_64-linux-gnu/libz.so /usr/lib/x86_64-linux-gnu/libaec.so /usr/lib/x86_64-linux-gnu/libm.so /usr/lib/x86_64-linux-gnu/libopenjp2.so /usr/lib/x86_64-linux-gnu/libexpat.so debug optimized /usr/lib/x86_64-linux-gnu/libnetcdf.so /usr/lib/x86_64-linux-gnu/hdf5/serial/libhdf5.so /usr/lib/x86_64-linux-gnu/libpthread.so
[16:17] <LocutusOfBorg> /usr/lib/x86_64-linux-gnu/libsz.so /usr/lib/x86_64-linux-gnu/libdl.so /usr/lib/x86_64-linux-gnu/libproj.so -lpangocairo-1.0 -lpango-1.0 -lgobject-2.0 -lglib-2.0 -lcairo Odb_fortran Odb eckit eckit_geometry eckit_linalg eckit_maths eckit_web eckit_cmd metkit -L/usr/lib/python3.7/config-3.7m-x86_64-linux-gnu -L/usr/lib -lpython3.7m -lpthread -ldl  -lutil -lm  -Xlinker -export-dynamic -Wl,-O1 -Wl,-Bsymbolic-functions]
[16:18] <LocutusOfBorg> but TBH, magics++ should add it as runtime dependency, because it is not direct to metview
[16:18] -queuebot:#ubuntu-release- Unapproved: accepted magics++ [sync] (disco-proposed) [3.2.1-1]
[16:18] <LocutusOfBorg> it won't probably fix the issue, but if not, I'll upload a magics++ with the runtime dependency added and open a new debian bug
[16:18] <LocutusOfBorg> (and understand why debian is good)
[16:19] <infinity> LocutusOfBorg: Yeah, I'd rather understand why the build works in debian before going and blindly hitting things with hammers.
[16:21] <LocutusOfBorg> I would say probably because of python3-default changes, but sure, I have to know why before complaining
[16:22] <LocutusOfBorg> just to avoid double work, once I fixed this, I plan to: gnu-smalltalk fix on ppc64el, seems just a CFLAG overridden but I failed so far
[16:22] <LocutusOfBorg> linker tries to merge libraries compiled with mlong-double-64 and libraries compiled without
[16:22] <LocutusOfBorg> so a matter of a configure.ac change, but I failed at it
[16:23] <infinity> LocutusOfBorg: If you have an idea how to fix that, go nuts.  I was just going to remove gnu-smalltalk in ppc64el to move past it.
[16:24] <LocutusOfBorg> does the explanation I did above make sense wrt linker?
[16:24] <LocutusOfBorg>  /usr/bin/ld: .libs/gstpub.o uses 64-bit long double, ../snprintfv/snprintfv/.libs/libsnprintfvc.a(libsnprintfvc_la-format.o) uses 128-bit long double
[16:25] <infinity> LocutusOfBorg: I mean, the bug (mixing ldbl-64 and ibm-ldbl-128) seems clear from the build log, but I wasn't aware there were compiler options to let it do it anyay.
[16:25] <LocutusOfBorg> don't forget I'm just a monkey typing random buttoms on a keyboard
[16:25] <infinity> s/anyay/anyway/
[16:25] <LocutusOfBorg> I think forcing snprintfvc to go 64bit is good, or should I move gstpub to 128?
[16:25] <LocutusOfBorg> the latter might be trivial, but I think they wanted to avoid it with introduction of this compiler option
[16:26] <LocutusOfBorg> seems something ppc64 specific code, so they might had good reasons to do it
[16:26] <LocutusOfBorg> after all, we still have days of autopkgtests ahead, so I plan to fix it properly
[16:26] <LocutusOfBorg> (both of them)
[16:26] <infinity> Erm, the systems libraried are ibm-ldbl-128.
[16:26] <infinity> libraries, even.
[16:27] <infinity> -mabi=ibmlongdouble might fix it.
[16:28] <LocutusOfBorg> confirmed, metview builds in a chroot with python3.7-dev installed
[16:28] <LocutusOfBorg> let me try debian now
[16:28] <infinity> Actually, I need to double check if we switched from IBM to IEEE when I wasn't looking.
[16:28] <infinity> But I tihnk that's scheduled for next year.
[16:28] <infinity> Oh!
[16:28] <infinity> -mlong-double-64 is on the commandline!
[16:29] <LocutusOfBorg> yes, what I said above :)
[16:29] <infinity> Dropping that entirely would probably cause gcc to just use the system default and not screw up.
[16:29] <LocutusOfBorg> upstream might have had good reasons to introduce it?
[16:29] <infinity> I doubt it.
[16:29] <LocutusOfBorg> lol, you don't usually trust upstream :)
[16:29] <infinity> Not as a general rule, no.
[16:29] <LocutusOfBorg> I know :D
[16:29] <infinity> Most upstreams (with a few exceptions) develop on one or two arches.
[16:30] <infinity> And think forcing compiler flags to conform to their arch is smarter than trusting system defaults to just be correct.
[16:30] <infinity> Note the warnings all over the logs too:
[16:30] <infinity> sinl.c:40:3: warning: floating constant exceeds range of ‘long double’ [-Woverflow]
[16:31] <infinity> Which implies their forcing 64-bit long doubles is exploding because we have 128-bit values coming from elsewhere.
[16:31] <LocutusOfBorg> ok, metview tries to link python3.7m in debian too, and the current build depedencies do *not* install it
[16:31] <LocutusOfBorg> how can the build succeed there?
[16:31] <LocutusOfBorg> https://buildd.debian.org/status/fetch.php?pkg=metview&arch=i386&ver=5.2.1-1&stamp=1540571808&raw=0
[16:32] <infinity> I assume -lpython3.7m came from elsewhere in Debian for some reason.
[16:32] <LocutusOfBorg> no, the link link has -lpython3.7m on both debian and ubuntu
[16:32] <infinity> I looked at that log and had the same WTF moment.
[16:32] <infinity> No, I mean, the .so is shipped in another package in Debian.
[16:32] <infinity> It's clearly there, or the link would be failing.
[16:32] <LocutusOfBorg> infinity, I installed the build deps in a clean chroot, and answer is no.
[16:33] <infinity> Also, this was 9 days ago.
[16:33] <LocutusOfBorg> infinity, I have a better bet: new magics++ doesn't really require it, so the linker is stripping
[16:33] <infinity> So... Rewind? :)
[16:33] <LocutusOfBorg> this is why I sycn'd it
[16:33] <infinity> "the linker is stripping"?
[16:33] <infinity> That's not how ld works.
[16:33] <LocutusOfBorg> ws-asneeded?
[16:33] <infinity> Debian doesn't build with as-needed.
[16:34] <infinity> And even if it did, -lfoo will always search for foo.so, then drop it on the floor if not needed.
[16:34] <LocutusOfBorg> mm this is also true...
[16:34] <LocutusOfBorg> find / -name "*python3.7m*so*"
[16:34] <LocutusOfBorg> nothing returned :/
[16:36] <infinity> Anyhow, the s390x build of magics++ is almost done, let's see in a bit if this magically (hah) fixes it and, if so, purge state from our brains and stop caring.
[16:37] <infinity> Cause I dunno about you, but I have more important things to worry about than wondering how the heck this is working.
[16:37] <LocutusOfBorg> I do have them :)
[16:37] <LocutusOfBorg> but http://debomatic-amd64.debian.net/distribution#unstable/magics++/3.2.1-1/buildlog
[16:37] <LocutusOfBorg> http://debomatic-amd64.debian.net/distribution#unstable/metview/5.2.1-1/buildlog
[16:37] <LocutusOfBorg> I also have two debian rebuilds ongoing
[16:46] <coreycb> doko: i'd like to upload a new version of nova with a small patch update if possible to disco in order to get an SRU under way. would that be ok?
[16:52] <infinity> coreycb: Upload away.  We might not accpet it if it'll get entangled in current transitions, but we'll accept it later.
[16:53] <coreycb> infinity: ok thanks
[16:55] -queuebot:#ubuntu-release- Unapproved: nova (bionic-proposed/main) [2:17.0.6-0ubuntu1 => 2:17.0.6-0ubuntu2] (openstack, ubuntu-server)
[16:55] -queuebot:#ubuntu-release- Unapproved: nova (disco-proposed/main) [2:18.0.1-0ubuntu1 => 2:18.0.1-0ubuntu2] (openstack, ubuntu-server)
[16:55] -queuebot:#ubuntu-release- Unapproved: nova (cosmic-proposed/main) [2:18.0.1-0ubuntu1 => 2:18.0.1-0ubuntu1.1] (openstack, ubuntu-server)
[16:58] <infinity> LocutusOfBorg: My Debian test build succeeded.   I think I need to do that again and ask sbuild to keep the chroot so I can figure out WTF.
[16:58] <infinity> (for metview)
[17:16] -queuebot:#ubuntu-release- Unapproved: libreoffice (disco-proposed/main) [1:6.1.2-0ubuntu4 => 1:6.1.3-0ubuntu2] (kubuntu, ubuntu-desktop)
[17:23] -queuebot:#ubuntu-release- Unapproved: odb-api (disco-proposed/universe) [0.18.0-5build1 => 0.18.0-6] (no packageset) (sync)
[17:29] <infinity> LocutusOfBorg: Okay, the difference is that the Debian build is installing libpython3.7, while the Ubuntu build isn't, and that's where the .so symlink lives.
[17:29] <infinity> Not entirely sure why...
[17:30] <infinity> libmagplus3v5 on Debian depends on libpython3.7
[17:30] -queuebot:#ubuntu-release- Unapproved: firefox (disco-proposed/main) [63.0+build2-0ubuntu0.18.10.2 => 63.0.1+build4-0ubuntu1] (kubuntu, mozilla, ubuntu-desktop)
[17:30] <infinity> It might not on Ubuntu due to as-needed?
[17:34] <infinity> Yep, from the Debian build log:
[17:34] <infinity> dpkg-shlibdeps: warning: package could avoid a useless dependency if debian/libmagplus3v5/usr/lib/s390x-linux-gnu/libMagPlusDouble.so.3.0.0 debian/libmagplus3v5/usr/lib/s390x-linux-gnu/libMagPlus.so.3.0.0 debian/libmagplus3v5/usr/lib/s390x-linux-gnu/libMagPlusSingle.so.3.0.0 were not linked against libpython3.7m.so.1.0 (they use none of the library's symbols)
[17:35] <infinity> LocutusOfBorg: Adding a build-dep on python-all-dev (to metview) is probably the path of least resistant for now.
[17:36] <infinity> LocutusOfBorg: But the more correct upstream solution for magics++ would be to stop overlinking (it links a TON of libs it doesn't seem to use) and then stop exporting those in its linker lines.
[17:37] -queuebot:#ubuntu-release- New: accepted linux-signed-azure [amd64] (bionic-proposed) [4.15.0-1031.32]
[18:17] <bdmurray> infinity: How do you feel about a speedy release of ubuntu-release-upgrader SRU? Bug 1799710 is blocking some upgrades to C and D.
[18:23] <infinity> bdmurray: What about the traceback in Comment 11?  You didn't respond to that on the bug.
[18:28] <bdmurray> infinity: I'll respond its update-manager and a known issue
[18:28] <infinity> Kay.
[18:29] <infinity> The patches look straightforward enough to me, I'm okay with expediting.
[18:29] <infinity> bdmurray: Did this bug not exist in older releases too?
[18:29] <infinity> (ie: does this affect xenial->bionic or trusty->xenial and if not, why not?)
[18:30] <infinity> Was bionic the first one to demand updates be installed?
[18:30] <bdmurray> Yes, that. I do want to get that into 16.04 though.
[18:30]  * infinity nods.
[18:30] <infinity> Releasing.
[18:30] <bdmurray> This improves the fix for bug 1797209
[18:31] <infinity> bdmurray: Don't forget to update meta-release for cosmic, this is the first SRU there.
[18:32] <bdmurray> infinity: The fix is in do-release-upgrade which doesn't come from the tarball, but its still a good idea and I'll do that.
[18:32] <infinity> Sure, I realise the fix is local, but people should always be getting the latest tarball.
[18:33] -queuebot:#ubuntu-release- Unapproved: sosreport (disco-proposed/main) [3.5-1ubuntu4 => 3.6-1] (ubuntu-desktop, ubuntu-server) (sync)
[18:34] -queuebot:#ubuntu-release- Unapproved: sosreport (disco-proposed/main) [3.5-1ubuntu4 => 3.6-1] (ubuntu-desktop, ubuntu-server) (sync)
[18:34] <infinity> bdmurray: Somewhat tempted to make "copy ubuntu-release-upgrade from release pocket to updates" part of the release process, and we could have meta-release always reference updates from release day.  Just so we don't forget that it's a knob that needs twiddling as some random later date.
[18:35] <infinity> s/as some/at some/
[18:35] <vorlon> xnox: it looks like perl on s390x is still unhappy despite your retrigger.  Did you get any further with this?
[18:35] <bdmurray> infinity: that seems reasonable to me
[18:36] <infinity> vorlon: To be fair, it's not "unhappy on s390x" so much as "autopkgtest is doing silly things on upgrade on all arches, but only s390x breaks because of the bootloader dep".
[18:37] <infinity> vorlon: See an amd64 log: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-disco/disco/amd64/a/alice/20181103_193821_5562e@/log.gz
[18:37] <infinity> vorlon: The following packages will be REMOVED:
[18:37] <infinity> vorlon: That's very much not what we want from the chroot upgrade, surely.
[18:37] <vorlon> sure
[18:38] <vorlon> is the fact that it fails a bug in the zipl package for not properly handling removed-not-purged?
[18:38] <infinity> Also probably yes.
[18:39] <vorlon> as for fixing the underlying failure, this seems to be the same as the problem xnox has hit before with systemd during autopkgtests, I have no idea what's going wrong with the upgrader selection :P
[18:40] <infinity> zz-update-grub starts with 'which update-grub >/dev/null 2>&1 || exit 0
[18:40] <infinity> zz-zipl probably wants a similar guard.
[18:41] -queuebot:#ubuntu-release- Unapproved: eccodes (disco-proposed/universe) [2.9.0-1 => 2.9.0-1ubuntu1] (no packageset)
[18:41] <infinity> I can JFDI that change.
[18:41] <infinity> xnox: ^
[18:41] <LocutusOfBorg> infinity, ^^ I think I fixed the eccodes too
[18:41] <LocutusOfBorg> I'll be afk, and finish magick later today if I don't fall aslee
[18:41] <LocutusOfBorg> *p
[18:41]  * LocutusOfBorg cheers!
[18:41] <infinity> LocutusOfBorg: I didn't know it was broken. :)
[18:42] <infinity> LocutusOfBorg: Oh, gross.
[18:42] <infinity> Does that maybe want something more dynamic?
[18:43] <infinity> Or just /usr/bin/python3 ... Why is it so specific?
[18:47] -queuebot:#ubuntu-release- Unapproved: s390-tools (disco-proposed/main) [2.6.0-0ubuntu7 => 2.6.0-0ubuntu8] (core)
[18:47] <infinity> vorlon: ^^
[18:47] <vorlon> looking
[18:47] -queuebot:#ubuntu-release- Unapproved: accepted eccodes [source] (disco-proposed) [2.9.0-1ubuntu1]
[18:48] <infinity> vorlon: Direct cargo-cult from grub, just s/update-grub/zipl/
[18:49]  * infinity goes for lunch.
[18:50] -queuebot:#ubuntu-release- Unapproved: accepted s390-tools [source] (disco-proposed) [2.6.0-0ubuntu8]
[18:50] -queuebot:#ubuntu-release- New sync: rust-mio (disco-proposed/primary) [0.6.16-1]
[18:54] <infinity> vorlon: s390-tools should probably migrate fairly quickly, and then I guess we can just retrigger $world on s390x and it'll still be done before everyone else. :P
[19:01] -queuebot:#ubuntu-release- Unapproved: libreoffice (bionic-proposed/main) [1:6.0.6-0ubuntu0.18.04.1 => 1:6.0.7-0ubuntu0.18.04.1] (kubuntu, ubuntu-desktop)
[19:07] -queuebot:#ubuntu-release- New: accepted rust-mio [sync] (disco-proposed) [0.6.16-1]
[19:09] -queuebot:#ubuntu-release- Unapproved: libreoffice-l10n (bionic-proposed/main) [1:6.0.6-0ubuntu0.18.04.1 => 1:6.0.7-0ubuntu0.18.04.1] (ubuntu-desktop)
[19:21] -queuebot:#ubuntu-release- New sync: rust-term-size (disco-proposed/primary) [0.3.1-2]
[19:28] -queuebot:#ubuntu-release- New sync: rust-nodrop-union (disco-proposed/primary) [0.1.9-1]
[19:28] -queuebot:#ubuntu-release- New sync: rust-rayon (disco-proposed/primary) [1.0.2-1]
[19:29] -queuebot:#ubuntu-release- New: accepted rust-nodrop-union [sync] (disco-proposed) [0.1.9-1]
[19:29] -queuebot:#ubuntu-release- New: accepted rust-rayon [sync] (disco-proposed) [1.0.2-1]
[19:31] -queuebot:#ubuntu-release- New binary: rust-nodrop-union [ppc64el] (disco-proposed/none) [0.1.9-1] (no packageset)
[19:31] -queuebot:#ubuntu-release- New binary: rust-nodrop-union [s390x] (disco-proposed/none) [0.1.9-1] (no packageset)
[19:31] -queuebot:#ubuntu-release- Unapproved: accepted livecd-rootfs [source] (disco-proposed) [2.543]
[19:32] -queuebot:#ubuntu-release- New binary: rust-nodrop-union [amd64] (disco-proposed/none) [0.1.9-1] (no packageset)
[19:33] -queuebot:#ubuntu-release- New binary: rust-nodrop-union [i386] (disco-proposed/none) [0.1.9-1] (no packageset)
[19:37] -queuebot:#ubuntu-release- New binary: rust-nodrop-union [arm64] (disco-proposed/none) [0.1.9-1] (no packageset)
[19:39] -queuebot:#ubuntu-release- New binary: rust-nodrop-union [armhf] (disco-proposed/universe) [0.1.9-1] (no packageset)
[19:39] -queuebot:#ubuntu-release- Unapproved: ironic (disco-proposed/universe) [1:11.1.0-0ubuntu5 => 1:11.1.0-0ubuntu6] (openstack)
[20:00] -queuebot:#ubuntu-release- Unapproved: rejected ironic [source] (disco-proposed) [1:11.1.0-0ubuntu6]
[20:01] <bdmurray> vorlon: Could you review the apport in the -bionic queue since you looked at the cosmic one?
[20:18] -queuebot:#ubuntu-release- Unapproved: grub2 (bionic-proposed/main) [2.02-2ubuntu8.8 => 2.02-2ubuntu8.9] (core)
[20:25] -queuebot:#ubuntu-release- Unapproved: grub2-signed (bionic-proposed/main) [1.93.9 => 1.93.10] (core)
[20:28] <vorlon> bdmurray: done
[20:29] -queuebot:#ubuntu-release- Unapproved: accepted apport [source] (bionic-proposed) [2.20.9-0ubuntu7.5]
[20:33] -queuebot:#ubuntu-release- Unapproved: net-snmp (disco-proposed/main) [5.7.3+dfsg-1.8ubuntu3.18.10.2 => 5.7.3+dfsg-1.8ubuntu4] (desktop-core, ubuntu-server)
[20:34] -queuebot:#ubuntu-release- Unapproved: accepted net-snmp [source] (disco-proposed) [5.7.3+dfsg-1.8ubuntu4]
[20:54] -queuebot:#ubuntu-release- New: accepted rust-nodrop-union [amd64] (disco-proposed) [0.1.9-1]
[20:54] -queuebot:#ubuntu-release- New: accepted rust-nodrop-union [armhf] (disco-proposed) [0.1.9-1]
[20:54] -queuebot:#ubuntu-release- New: accepted rust-nodrop-union [ppc64el] (disco-proposed) [0.1.9-1]
[20:54] -queuebot:#ubuntu-release- New: accepted rust-term-size [sync] (disco-proposed) [0.3.1-2]
[20:54] -queuebot:#ubuntu-release- New: accepted rust-nodrop-union [arm64] (disco-proposed) [0.1.9-1]
[20:54] -queuebot:#ubuntu-release- New: accepted rust-nodrop-union [s390x] (disco-proposed) [0.1.9-1]
[20:54] -queuebot:#ubuntu-release- New: accepted rust-nodrop-union [i386] (disco-proposed) [0.1.9-1]
[20:55] -queuebot:#ubuntu-release- New binary: rust-term-size [s390x] (disco-proposed/none) [0.3.1-2] (no packageset)
[20:56] -queuebot:#ubuntu-release- New binary: rust-term-size [arm64] (disco-proposed/none) [0.3.1-2] (no packageset)
[20:57] -queuebot:#ubuntu-release- New binary: rust-term-size [amd64] (disco-proposed/none) [0.3.1-2] (no packageset)
[20:57] -queuebot:#ubuntu-release- New binary: rust-term-size [armhf] (disco-proposed/none) [0.3.1-2] (no packageset)
[20:57] -queuebot:#ubuntu-release- New binary: rust-term-size [i386] (disco-proposed/none) [0.3.1-2] (no packageset)
[20:58] <bdmurray> vorlon: thanks!
[20:58] -queuebot:#ubuntu-release- New binary: rust-term-size [ppc64el] (disco-proposed/none) [0.3.1-2] (no packageset)
[21:01] -queuebot:#ubuntu-release- New: accepted rust-term-size [amd64] (disco-proposed) [0.3.1-2]
[21:01] -queuebot:#ubuntu-release- New: accepted rust-term-size [armhf] (disco-proposed) [0.3.1-2]
[21:01] -queuebot:#ubuntu-release- New: accepted rust-term-size [ppc64el] (disco-proposed) [0.3.1-2]
[21:01] -queuebot:#ubuntu-release- New: accepted rust-term-size [arm64] (disco-proposed) [0.3.1-2]
[21:01] -queuebot:#ubuntu-release- New: accepted rust-term-size [s390x] (disco-proposed) [0.3.1-2]
[21:01] -queuebot:#ubuntu-release- New: accepted rust-term-size [i386] (disco-proposed) [0.3.1-2]
[21:15] <vorlon> Depends: s390-tools net-snmp (not considered)
[21:15] <vorlon> infinity: ^^ bwahaha kill me now
[21:16] <vorlon> looks suspiciously like net-snmp has idiotic shlibdeps
[21:18] <vorlon> infinity: oh looks like you're already on top of net-snmp
[21:18] <vorlon> infinity: are you reuploading s390-tools also?
[21:30] -queuebot:#ubuntu-release- Unapproved: s390-tools (disco-proposed/main) [2.6.0-0ubuntu8 => 2.6.0-0ubuntu9] (core)
[21:30] <vorlon> infinity: ^^
[21:44] -queuebot:#ubuntu-release- New sync: rust-rayon-core (disco-proposed/primary) [1.4.1-1]
[21:44] -queuebot:#ubuntu-release- New: accepted rust-rayon-core [sync] (disco-proposed) [1.4.1-1]
[21:46] -queuebot:#ubuntu-release- New binary: rust-lazycell [s390x] (disco-proposed/none) [1.2.0-1] (no packageset)
[21:47] -queuebot:#ubuntu-release- New binary: rust-lazycell [amd64] (disco-proposed/none) [1.2.0-1] (no packageset)
[21:48] -queuebot:#ubuntu-release- New binary: rust-lazycell [i386] (disco-proposed/universe) [1.2.0-1] (no packageset)
[21:52] -queuebot:#ubuntu-release- New binary: rust-lazycell [ppc64el] (disco-proposed/universe) [1.2.0-1] (no packageset)
[21:56] <infinity> vorlon: Oh, I had one pending upload, but was on the phone.  I'll take yours instead.
[21:57] -queuebot:#ubuntu-release- Unapproved: accepted s390-tools [source] (disco-proposed) [2.6.0-0ubuntu9]
[21:57] -queuebot:#ubuntu-release- New binary: rust-lazycell [arm64] (disco-proposed/universe) [1.2.0-1] (no packageset)
[22:01] -queuebot:#ubuntu-release- New binary: rust-lazycell [armhf] (disco-proposed/universe) [1.2.0-1] (no packageset)
[22:13] -queuebot:#ubuntu-release- Unapproved: python-pysnmp4 (disco-proposed/main) [4.4.6-1 => 4.4.6+repack1-1] (ubuntu-server) (sync)
[22:18] -queuebot:#ubuntu-release- New: accepted rust-lazycell [armhf] (disco-proposed) [1.2.0-1]
[22:19] -queuebot:#ubuntu-release- New: accepted rust-lazycell [amd64] (disco-proposed) [1.2.0-1]
[22:19] -queuebot:#ubuntu-release- New: accepted rust-lazycell [i386] (disco-proposed) [1.2.0-1]
[22:19] -queuebot:#ubuntu-release- New: accepted rust-lazycell [arm64] (disco-proposed) [1.2.0-1]
[22:19] -queuebot:#ubuntu-release- New: accepted rust-lazycell [ppc64el] (disco-proposed) [1.2.0-1]
[22:19] -queuebot:#ubuntu-release- New: accepted rust-lazycell [s390x] (disco-proposed) [1.2.0-1]
[22:51]  * vorlon queues up a bunch of autopkgtest retries for perl/s390x
[22:52] <bdmurray> tsimonq2: The new ubuntu-release-upgrader was released.
[23:01] -queuebot:#ubuntu-release- Unapproved: accepted tmux [source] (bionic-proposed) [2.6-3ubuntu0.1]
[23:05] -queuebot:#ubuntu-release- Unapproved: rejected python-tornado [source] (bionic-proposed) [4.5.3-1.0ubuntu1]
[23:07] -queuebot:#ubuntu-release- Unapproved: rejected mutter [source] (bionic-proposed) [3.28.3-2~ubuntu18.04.2]
[23:11] -queuebot:#ubuntu-release- Unapproved: rust-crossbeam-utils (disco-proposed/universe) [0.4.1-1 => 0.5.0-1] (no packageset) (sync)
[23:58] <mwhudson> oh ffs why run tests during the build and ignore the results