[00:50] flocculant: yes, this is a very ugly bug [01:27] cyphermox: I can't work out from the code where the formatted strings are coming from, do you know? [01:33] that's what's terrible [01:34] slangasek: looks like they come from apt, which has always been outputting things in a sprintf using '%.4f', but now is setting locale more forcefully [01:34] right, I assumed apt but couldn't find where in the python code apt was being invoked :) [01:34] unfortunately, this means including LC_NUMERIC; so your 0.0000 in .4f becomes 0,0000 in fr and es, for example [01:35] it's sometimes python-apt, sometimes called through debconf-apt-progress [01:36] master bug is bug 1611010 [01:36] bug 1611010 in apt (Ubuntu) "yakkety desktop - non-english installation crashes with /plugininstall.py: ValueError: invalid literal for int() with base 10: ''" [Critical,Confirmed] https://launchpad.net/bugs/1611010 [01:42] slangasek: I just wasn't prepared to patch apt directly before talking to juliank or someone else about it, but looks like it ought be be fixed there. [02:54] please try a rebuild of Ubuntu Studio's yakkety iso's, germinate does not show gnupg1 now [04:57] slangasek: robru: that force-merge was related to the transition and the need to do next landing to OTA-13 for the affected package. it was then added to our backlog list of "monitor migration to release pocket manually", which is why we asked for the removal of the s390x binaries now. [04:58] (the backlog list is now quite short http://pad.ubuntu.com/yakkety-pending-landings ) [04:58] Mirv: thanks for clarifying [04:58] oh, and sync-monitor migrated, great [04:59] and someone did the online-accounts too eventually so I think that list might be down to zero, let's see [05:00] yeah, I did that one when I was looking through update_excuses output for other stale s390x binaries; those were the only two I saw [05:01] slangasek: yeah I asked pitti for it yesterday, he said he did it but apparently something went wrong and they weren't removed. anyway, it unblocked the four pending packages, while sync-monitor was fifth so now the backlog is cleared! thanks! [05:02] hmm, I did this: [05:02] remove-package -s yakkety-proposed -b -a s390x -m 'dependencies got removed on s390x' ubuntu-system-settings-online-accounts libonline-accounts-client1 libonline-accounts-client-dev qtdeclarative5-online-accounts-client0.1 qml-module-ubuntu-onlineaccounts-client qml-module-ubuntu-onlineaccounts-client-doc libonline-accounts-plugin-dev ubuntu-system-settings-online-accounts-autopilot [05:02] (from my bash history) [05:02] LP didn't raise an error, was that wrong or was there a fluke? [05:04] sounds correct, it just didn't seem to happen according to excuses page or when I checked https://launchpad.net/ubuntu/yakkety/s390x/ubuntu-system-settings-online-accounts [05:04] hmm [05:06] pitti: fwiw, your run doesn't show here: https://launchpad.net/ubuntu/yakkety/s390x/libonline-accounts-plugin-dev/ [05:06] maybe you forgot to 'y' it? [05:06] (or LP bug) [05:07] oh, also britney did the thing again where the binaries I just removed from -proposed were copied to yakkety, sigh [05:26] slangasek: I guess it was me then, or some fluke; I *do* see copy-package be ignored very often (i. e. sru-release, need to run it twice way too often), but I never saw remove-package silently fail that way [05:26] so I guess I screwed up something [05:26] but good to know for next time that this the right command in general [05:38] Mirv: could you check what's needed for the evolution 3.22 transition to finish? qtorganizer5-eds/s390x at least is a problem [05:39] jbicha: well we have pitti here now so maybe he could remove the s390x for it: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#qtorganizer5-eds - at least manually looking at the excuses page that would seem the only one. I will just need to double check if qtorganizer5-eds has been fixed to not bring back the s390x binaries accidentally. [05:42] so yes indeed it doesn't have the binaries and shouldn't be able to bring them back either. the s390x binaries should be removed from the release pocket version of qtorganizer5-eds. [06:49] slangasek: sorry - if I'd looked a bit further down the ubiquity bug list I might have seen the one jibel had commented on and not bothered you ;) [06:50] cyphermox: yep - not one I'd have seen if one of my testers hadn't said anything to be - ignorance is bliss sometimes === agateau_ is now known as agateau === ant_ is now known as Guest15611 [08:50] greetings, the Studio iso failed again. 'kactivitymanagerd : Breaks: kactivities (< 5.20~) but 5.18.0-0ubuntu1 is to be installed' [08:59] not sure i understand the germinate link jbicha gave me right, but it seems to be plasma framework and calligra that need kactivities.. [09:00] i have to be afk for an hour now unfortunately.. but will read backlog of course [09:11] sakrecoer: libkactivities6 has a Recommends on kactivities, which is a package that kactivities-kf5 has dropped [09:12] yofel: ^- you might be interested in that knowledge === pete-woods[fail] is now known as pete-woods [10:41] thank you Laney :) [10:42] sakrecoer: #kubuntu-devel is a good place to ask about that [10:43] Laney: i will do reach out to them, thank you! [12:51] please remove qtorganizer5-eds/s390x [14:08] slangasek, hey could you do another favor? Could you remove address-book-app from yakkety/s390 to unblock this silo: https://requests.ci-train.ubuntu.com/#/ticket/1815 [14:38] renatu: there are no address-book-app binaries on s390x in yakkety; that silo says "needs a rebuild"? [14:39] slangasek, hooo, this changed now. Let me re-build it [14:39] thanks [14:40] jbicha: qtorganizer5-eds/s390x removed [14:49] slangasek: we also need indicator-network s390x deleted to unstick http://people.canonical.com/~ubuntu-archive/proposed-migration/yakkety/update_excuses.html#indicator-network thanks [15:18] robru: done [15:18] slangasek: thanks! === dgadomski_ is now known as dgadomski [15:38] michi, michi_: hi, would you rebuild storage-framework against the current boost (1.61) please? i think a no-change rebuild should do it [15:41] question: why can't vcmi migrate to yakkety release? [15:41] old binaries left on amd64: vcmi-dbg (from 0.98+dfsg-2.1ubuntu2) [15:42] isn't this stuff autodecrufted magically? [15:43] not if it's NBS in proposed [15:46] what does it mean? [15:46] I did the ddbgsym migration stuff [15:46] and in fact I see produced: vcmi 0.98+dfsg-3 and vcmi-dbgsym 0.98+dfsg-3 [15:46] of course I dropped the -dbg package... [15:47] actually there's no automatic removal anywhere, but if you drop something and there's a binary left over in -proposed then it'll hold up the migration like this [15:47] this is the commit https://anonscm.debian.org/cgit/pkg-games/vcmi.git/commit/?id=19efe46c326202a2de9830609b26961a6ea48d69 [15:47] If it's only left over in release then it doesn't, and just gets picked up in the NBS report for someone to act on later [15:47] so, what Debian does automagically is done manually in Ubuntu? [15:47] basically you need someone to delete that binary for you [15:48] because Debian is removing the -dbg packages automatically on autodecruft AFAIK [15:48] I never saw a migration hold because of a dbg package missing [15:49] michi, michi_: nvm, i think that was a false positive because of what is in control, i looks like it actually built against the current version [15:52] any ~ubuntu-archive around? please decruft supercollider-supernova and remove boost1.58 [15:54] sakrecoer: Are you my release buddy for beta 1 or is that someone else? [15:55] I'm setting up the stuff now - freeze, ISO tracker milestone, stopping automated builds [15:55] Laney: not sure he's about currently - but yes he is and me in the background helping him out where necessary [15:55] ok [15:56] flocculant: I see: studio, GNOME, kubuntu, lubuntu - correct? [15:56] afaik yes [15:57] though I am surprised Mate aren't [15:57] flexiondotorg: are you in? [15:57] Laney: from an old mail "For completeness Ubuntu MATE intend to participate in Alpha 1, Alpha 2, Beta 1 and final beta." [15:58] cool, thanks [15:58] welcome :) [15:58] tsimonq2: Do you want powerpc this time? [15:58] Laney: yeah, i am :) [15:58] Laney: obviously currently studio won't build - and I assume kubuntu won't either [15:59] well, i'm guided by tsimonq2 :) [15:59] sakrecoer: aah - didn't know you were awake :D [15:59] Laney: Lubuntu doesn't do PPC for non-LTS [15:59] ok [15:59] infinity, has the mark packages in main as 5year support happened in xenial 16.04.1 [15:59] flocculant: Kubuntu seems okay so far https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/xenial/kubuntu [15:59] they must seed some different stuff [16:00] LocutusOfBorg: when a binary is in devel and is dropped from the version of package in devel-proposed, proposed-migration figures it out on its own. But when a new version is uploaded to devel-proposed, and there was a previous unmigrated version already in devel-proposed that didn't migrate yet and built a different set of binaries, the archive admins have to figure it out by hand [16:00] infinity, and/or do things need to be SRUed to get the 5 year marker? [16:00] Laney: studio iso still doesn't build though.. waiting for kubuntu user to join and help us with kactivitymanagerd [16:00] thanks slangasek [16:00] or shall i just seed the s390-tools? [16:00] LocutusOfBorg: (vcmi-dbg removed) [16:00] thanks [16:00] sakrecoer: That's fine: you have two days :P [16:00] Laney: don't you mean yakkety? https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/yakkety/kubuntu [16:00] (and yes, flocculant has been a huge help) [16:00] next time I'll do a fake upload removing the dbg and then a new upload force-syncd [16:00] :) [16:00] flocculant: oh right [16:01] unlucky then :) [16:01] cos studio and kubuntu are having fun with the same things :p [16:01] ginggs_: supercollider-supernova> what needs decrufted where? this isn't listed on update_excuses or nbs [16:01] then they also have two days to fix it [16:01] :) [16:01] * flocculant likes the 'we're only doing final beta stance' that we took :D [16:02] slangasek: reverse-depends src:boost1.58 - shows supercollider [16:02] ginggs_: and this was removed from Debian? [16:02] demote to proposed? [16:02] Laney, wrestling a large server deployment. [16:02] But post and I'll catch you in a bit. [16:02] slangasek: maybe i misread something, i thought these were old binaries left behind [16:02] xnox: hisss [16:03] slangasek, or maybe i should fix it. [16:03] slangasek, oh weait it looks like it depends on things only on some arches. smells fishy === Laney changed the topic of #ubuntu-release to: Released: Trusty 14.04.5, Xenial 16.04.1 | Archive: feature freeze, beta 1 freeze | Yakkety Release Coordination | Please don't upload things during freezes where you shouldn't, or be prepared to apologise to the release team | We accept payment in cash, check or beer | melior malum quod cognoscis [16:03] well what I currently see is that supercollider has different package versions across different archs within yakkety [16:03] so... yum? [16:03] slangasek:the supercollider-supernova binaries are only made for amd64, i386 and x32 now [16:04] yeah, so how did it get into yakkety before the old binaries were dropped [16:04] but supercollider is built on all arches in yakkety [16:04] Laney: this list of particiaption should be accurate https://wiki.ubuntu.com/YakketyYak/Beta1 [16:04] sakrecoer: thx [16:04] well, it's mentioned in my force-hint for the Qt transition, but that shouldn't have the same effect as 'force' [16:05] interesting use of latin :D [16:05] ah I see [16:05] it's half-NBS :) [16:06] http://people.canonical.com/~ubuntu-archive/architecture-mismatches.html looks strange too, but not for supercollider. [16:06] ginggs_: ok, removing [16:06] slangasek: thanks :) [16:08] storage-framework has only alternative deps [16:12] xnox, slangasek: so boost1.58 can go now? [16:12] ginggs_, i think so. [16:12] Oops [16:12] Don't know how to remove all those Kubuntu products [16:12] slangasek, filed bug #1616130 [16:12] ginggs_: yes, though I'm having a grump that 'remove-package' doesn't give me an option to remove only the source package and leave the binaries for NBS [16:13] bug 1616130 in boost1.58 (Ubuntu) "RM boost1.58, superseeded" [Undecided,New] https://launchpad.net/bugs/1616130 [16:13] but it will never be NBS, because there are alternative build-deps on boost1.58 [16:13] e.g. libboost-foo-dev | libboost1.58-foo-dev [16:13] (as in green in NBS) [16:14] xnox: it's NBS as soon as the source is gone ;) [16:14] and AFAIK nbs is smart enough to work out alt deps [16:16] I need an autopkgtest on s390x marked as okay instead of a regression: http://people.canonical.com/~ubuntu-archive/proposed-migration/yakkety/update_excuses.html#url-dispatcher [16:21] stgraber: for the lubuntu-next images, I assume an update to queuebot is needed to add those? [16:38] Laney, I'm available if you are. [16:38] flexiondotorg: Never mind [16:38] OK [16:44] yofel: hi! there's a whole bunch of kf/etc related autopkgtest issues again in excuses for qtbase-opensource-src and qtdeclarative-opensource-src, many but not all say "cc1: warning: command line option ‘-std=gnu++11’ is valid for C++/ObjC++ but not for C". I'm not sure how to organize getting the Qt updates in, would you like to check the issues and say "please override" to archive admins? [16:44] bluez-qt is the only clear easy fix (one symbol change) [16:45] the qtbase change specifically doesn't do anything but a configure change affecting powerpc only, so it's essentially a no-change rebuild [16:46] xnox: Hasn't happened yet, but no things don't need to be reuploaded. [16:46] ok. [16:46] infinity, any ETA, and/or may I seed the s390x packages pointlessly until that's done? [16:47] cause nobody actually cares about the bulk of 9m marked packages =) [16:48] xnox: Seeding things won't fix anything... [16:48] i thought it would, as in launchpad would republish xenial-updates whenever that moves, and it would have the new support stanzas. [16:49] the problem with s390-tools is that it's in main, without actually being seeded into anything. [16:49] xnox: Oh, it would "fix" updates, yes, but not the release pocket. [16:49] i / nobody cares about the release pocket either really =) [16:49] xnox: But is IBM not satisfied with "yes, it's supported, we'll fix the metadata soon"? [16:53] infinity, i'm pretty sure they do not care about release pocket metadata. they care for $ ubuntu-support-status to not show that s390-tools aka bootloader is supported only until January 2017 [16:53] and thus the metadata in the -updates [16:56] xnox: Well, we care about both, since we're not going to SRU all of main. [16:57] well yeah. let's pretend that release + -security is a valid thing, and yeah we'd need to fix release properly to show proper status. [16:57] and imho s390-tools should be seeded somewhere anyhow. [16:58] i guess it's a unicorn package that is in boot seed and nowhere else that is in use. [16:58] it's not like seeding it into supported will change anything, in any of the metas. [16:58] Sure, it should be in boot. [16:59] Which is is. [17:00] it is in boot¸ but boot is not included into supported, hence no 5y badge [17:00] Right, so that's what I'm fixing. [17:00] Stop trying to "fix" it with seeds. [17:01] I guess I can land fix 1 before fix 2 (ie: we can fix the updates pocket today and release later) [17:02] yeah. "regenerate release pocket" is out of scope for the request. [17:02] "regenerate release pocket" is part of separate request. [17:03] surely that is a fix in seed structure then? e.g. supported-installer-common: standard installer -> supported-installer-common: standard installer boot [17:05] oh, this would fix other bootloaders too. [17:05] because at the moment grub-efi-amd64-signed is similarly on 9m badge. === Guest25180 is now known as med_ === med_ is now known as medberry [17:06] but that would be broken as well. === medberry is now known as med_ [17:06] cause some hwe kernels are not 5yr. [17:06] i gave up. [17:25] oh, I guess evolution 3.22 was just a bit too late for beta 1 but I don't particularly need it in the beta so it'll be ok [17:32] tedg: surely indicator-session's autopkgtest is failing because it depends on url-dispatcher which has been removed on s390x, and we should remove indicator-session also? [17:44] slangasek, now is failing again: https://requests.ci-train.ubuntu.com/#/ticket/1815 [17:45] slangasek, this is the excuses file: https://requests.ci-train.ubuntu.com/static/britney/ticket-1815/landing-049-yakkety/excuses.html [17:47] renatu: that shows that there are old s390x binaries in the silo which need to be removed [17:47] "from 0.2+16.10.20160822.1-0ubuntu1" tells you that it's not the version that's in the archive [17:47] so any trainguard should be able to do this fwiw [17:47] slangasek, humm ok thanks [17:47] renatu: (but I've got this one) [17:48] slangasek, thanks [17:49] $ remove-package -m NBS -s yakkety -A ~ci-train-ppa-service/ubuntu/landing-049 -a s390x -b qtdeclarative5-ubuntu-addressbook0.1 qtdeclarative5-ubuntu-contacts0.1 [17:49] (done) [20:16] Laney: Ubuntu Studio need ubuntustudio-meta unblocked to hopefully fix Live CD build [20:34] tedg: have convinced myself that yes, indicator-session binary should be removed from s390x; done [20:37] slangasek: Great, thank you! [21:41] tsimonq2: nope, queuebot doesn't know anything about the products, it triggers on anything that's on the tracker [21:42] ok cool, thanks stgraber