[03:17] pitti: 1500193 is 14 hours without a retrace, are the retraces blocked by the recent lcy01 issues? === hikiko__ is now known as hikiko [06:29] Good morning [06:29] Hello, any kernel maintainer here, or a How To build *only one* flavor in a kernel package that is building via Launchpad (PPA) and configured via Git ? [06:29] sarnold: no, they just crash often because some ddeb download issues from LP; I'm hand-holding them [06:34] NikTh: Pretty sure you could delete whichever of debian.master/control.d/vars.{generic,generic-lpae,lowlatency} you didn't want. [06:36] RAOF: I'm missing something here. I have did what you say, but maybe I need to alter some scripts ? I don't know. Check (when you have time) a build fail here: https://launchpadlibrarian.net/218918782/buildlog_ubuntu-trusty-amd64.linux_4.2.1-999.2015Sep26_BUILDING.txt.gz [06:36] RAOF: My propose is to minimize the building time, by removing any not required flavors(generic-lowlatency..etc). [06:37] RAOF: I have removed some files from debian.master/config , debian.master/control.d/, also modified debian.master/rules.d/amd64 and i386 (skipabi, skipmodules, flavors) but apparently something is missing [06:38] Strange error, and I'm not deeply familiar with the kernel build system. [06:41] NikTh: It might be quicker to just not bother restricting flavours; the whole kit and caboodle is a ~1hr build... [06:41] RAOF: No prob , thanks for trying to help :) Any other suggestion is welcome. I'm testing various changes local with 'pbuilder' now. Still getting the same error. [06:43] NikTh: I'd probably edit debian/rules.d/0-common-vars.mk to set AUTOBUILD=true (which will mean you can drop most of your other changes). That and deleting the control.d/vars.$FOO might be all you need? [06:43] RAOF: 4-5 hours, not 1hr , from what I've seen. It's OK, packages are not building in my laptop but in a virtual launchpad builder, but it would be better if I could reduce this time and get rid of not required flavors (generic-lowlatency..etc) :) === kickinz1|eow is now known as kickinz1 [06:44] RAOF: I will try this ASAP. Thanks ! [07:02] good morning [07:03] hey dholbach [07:03] salut seb128 === athairus is now known as athairuzzz [07:52] Hello [07:52] I want to make some requests for you (for Ubuntu Willy [07:52] ) [07:55] Here I will send you a paste of my bash sesion trying to install Minetest [07:55] http://paste.ubuntu.com/12599522/ [07:55] and here what I got when trying to install libleveldb1 [07:55] http://paste.ubuntu.com/12600703/ [07:56] here we need to edit the Minetest packages to adapt them to the new dependencies names === Zic_ is now known as Zic === lan3y is now known as Laney === Laney is now known as Guest61035 [08:05] what do you say friends? [08:06] I think its important to keep all the packages present on the Ubuntu's official repos up to date not only on the app's version, even on the dependencies of the apps [08:06] no? === Guest61035 is now known as Laney [08:06] hello Laney [08:10] I saw something bad? [08:10] :( [08:11] I am trying to keep the Ubuntu's repo's packages up to date [08:11] that is all :) [08:28] dear admins [08:28] somebody could help me? [08:29] SergioEDuran1: Ubuntu's minetest has the correct dependency - does apt-cache policy minetest show it as coming from a PPA? === _morphis is now known as morphis [08:32] Laney it is from the official Ubuntu's repos [08:32] http://paste.ubuntu.com/12599522/ [08:32] No [08:32] http://paste.ubuntu.com/12600703/ [08:32] Depends: minetest-data (= 0.4.12+repack-2ubuntu1), libc6 (>= 2.15), libcurl3-gnutls (>= 7.16.2), libfreetype6 (>= 2.2.1), libgcc1 (>= 1:4.1.1), libirrlicht1.8, libjsoncpp0v5, libleveldb1v5, libluajit-5.1-2, libopenal1 (>= 1.14), libsqlite3-0 (>= 3.5.9), libstdc++6 (>= 5.2), libvorbisfile3 (>= 1.2.0), libx11-6, zlib1g (>= 1:1.1.4) [08:32] you can check the paste :) [08:33] I have not used any Minetest ppa [08:33] Maybe run the command I gave you so we can see which minetest it is trying to install [08:34] Ok [08:36] here it is Laney [08:36] http://paste.ubuntu.com/12600910/ [08:36] "apt-cache policy minetest" please [08:38] Ok [08:41] I saw the problem, solved :) [08:41] and why not packaging GNOME 2048? [08:41] Gnome bug 2048 in general "Miscount of no of emails in Clock and Mail Notify Applet" [Normal,Resolved: duplicate] http://bugzilla.gnome.org/show_bug.cgi?id=2048 [08:41] gnome-2048 [08:41] hehehehehe [08:42] and maybe adding gnome-2048 gnome-taquin and hitori to the gnome-games Laney [08:42] ? === larsu_ is now known as larsu [10:35] friends [10:35] I am here again [10:35] I have sleeped a little bit [10:36] so... what do you think about including GNOME 2048 on Ubuntu 15.10 as part of the GNOME games? [10:36] Gnome bug 2048 in general "Miscount of no of emails in Clock and Mail Notify Applet" [Normal,Resolved: duplicate] http://bugzilla.gnome.org/show_bug.cgi?id=2048 [11:26] infinity: wgrant: cjwatson: What's the canonical way of finding the architecture that I'm building for in a livecd-rootfs (binary) hook? [11:27] Odd_Bloke: $LB_ARCHITECTURES (why it's plural I'm not quite sure) [11:28] hm let me just check it gets that bit of environment [11:28] cjwatson: I don't _think_ it does. [11:29] Odd_Bloke: ok, just use dpkg --print-architecture then [11:29] Indeed, we use that in a few places in hooks [11:31] OK, cool. === hikiko is now known as hikiko|ln === lionel_ is now known as lionel === _salem is now known as salem_ [12:44] RAOF: Unfortunately this gave me the same error. I'm trying now same thing but in addition I have modified debian.master/rules.d/amd64.mk and i386.mk to exclude generic and lowlatency from 'Flavors = ' line. [12:46] Oh, and also I have remove config.flavor.generic and lowlatency from 'configs' folder. === hikiko-lpt is now known as hikiko [13:26] anyone knows if ubuntu implements cpu shielding (cgroup cpuset) with systemd? === gammax90 is now known as gammax === Guest62641 is now known as balloons [13:59] cyphermox, that has no working retrace but it's high ranked on wily e.u.c, https://errors.ubuntu.com/problem/e72162eedbbe2652de316ee8653c35ee9b3ddebe [13:59] so might be worth investigating [14:00] wpa_supplicant segfault, seems to have started with 2.4 [14:03] yeah, I know :( [14:03] just, no retrace and already complicated bugs with one [14:03] bdmurray, do you know why https://errors.ubuntu.com/problem/56d2e995b3803bec644316604eb9cad22270532c claims " [14:03] The following packages are missing debug symbols: libtrust-store2" where http://ddebs.ubuntu.com/pool/universe/t/trust-store/ seems to have those? === salem_ is now known as _salem === _salem is now known as salem_ === juliank0 is now known as juliank === ralsina_ is now known as ralsina [14:41] pitti: kcoreaddons looks blocked on regressions in knewstuff and khtml but there's no failures that I can see [14:44] Riddell: indeed; I'll deal with this [14:45] pitti: some similar ones, kio and ki18n on kdelibs4support [14:45] Riddell: kwin, ktexteditor, and kdepim-runtime look real to me, but your's indeed not [14:49] Riddell: I'll check again in a bit, most stuff shold be right now [14:53] seb128: I don't and will look into it some more. [14:53] bdmurray, thanks [14:54] RAOF: We got it :-) [14:54] RAOF: Thank for you valuable help ! [15:08] pitti: now for kcoreaddons it seems all is good but it still says not considered and I'm unsure why http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#kcoreaddons [15:09] pitti: also kwindowsystem has ktextwidgets as regressions but there is none http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#kwindowsystem [15:11] Riddell: yup, currently doing the next iteration (sorry, at sprint, not much bandwidth) [15:39] Riddell: so a big bunch now landed [15:39] Riddell: http://autopkgtest.ubuntu.com/packages/k/kdepim-runtime/wily/amd64/ seems real, the rest is fixed (or covered by your force-*) [15:40] pitti: kdepim-runtime just got uploaded so maybe that fixes it [15:40] Riddell: FYI, I by and large know what's wrong in britney wrt. the false "regression" results, will fix after the sprint [15:46] hey pitti, I need to backport some stuff from Debian's ifupdown (to fix bug #1337873). How can I offer the changes if the Ubuntu ifupdown package format does not support quilt? Is there Ubuntu git tree for ifupdown? [15:46] bug 1337873 in ifupdown (Ubuntu) "Precise, Trusty, Utopic - ifupdown initialization problems caused by race condition" [Undecided,In progress] https://launchpad.net/bugs/1337873 [15:46] dgadomski: it's a native package, just apply it inline [15:46] dgadomski: and attach debdiffs to the bug [15:46] pitti: alright, thanks! [15:47] slangasek, stgraber, So I just wanted to bring up again that we are looking for a vic...volunteer aa who could do the final parts to let dpdk (bug 1487538) into wily. Note that the PPA version linked in the bug is behind by 2 commits (visible in git) that arges made when sponsoring. [15:47] bug 1487538 in Ubuntu "Add dpdk to wily universe" [Wishlist,New] https://launchpad.net/bugs/1487538 [15:51] smb: I can probably take a look at that today, once my web browser decides it's also willing to come back from vacation [15:52] If someone can confirm this : bug 1500216 [15:52] bug 1500216 in unity-control-center (Ubuntu) "Installing/Updating language support failed with "org.freedesktop.DBus.Error.NoReply" message" [Undecided,New] https://launchpad.net/bugs/1500216 [15:53] slangasek, cool. thank you and good luck. things always break when one isn't looking [15:57] slangasek: I have a bit (understatement) of context on dpdk, but would be much happier if you did the NEW review, since you're picky about different things than I am. Feel free to waste time on our call today discussing it, though. [16:19] slangasek, infinity: openvswitch-dpdk is also in the queue, depends on dpdk - it will pass its autopkgtests but only if the default cpu level if bumped via qemu options... [16:19] needs sse3 [16:32] dpm, pitti, wgrant, gedit.mo seems to be missing from the current wily langpacks (at least for de and fr), do you have any idea why? [16:34] dpm, pitti, wgrant, looks fine the most recent update dropped quite some translations :-/ [16:35] $ debdiff language-pack-gnome-de_15.10+20150922_all.deb language-pack-gnome-de_15.10+20150924_all.deb [16:35] Files in first .deb but not in second [16:35] -rw-r--r-- root/root /usr/share/locale-langpack/de/LC_MESSAGES/eog.mo [16:35] -rw-r--r-- root/root /usr/share/locale-langpack/de/LC_MESSAGES/evolution-3.16.mo [16:35] -rw-r--r-- root/root /usr/share/locale-langpack/de/LC_MESSAGES/file-roller.mo [16:35] etc [16:36] seb128, no idea. Let me look at the template [16:37] dpm, thanks [16:38] seb128, hm, I cannot see anything obvious. Last template update was on 2015-08-12 [16:38] priority seems correct [16:38] dpm, there are like 10 of those which vanished in friday's update [16:40] I'm looking at the content of language packs to see if they are in the export that's used as the source of the packages [16:40] well,, downloading atm [16:41] same here [16:41] but they are 258M for the 22 and for the 24 [16:41] so it doesn't look like the content changed [16:41] well, I'm speaking about the tarballs export on https://translations.launchpad.net/ubuntu/wily/+language-packs [16:42] yeah [16:45] pitti: oh, that's a shame about the retracers; I hadn't noticed any issues in a long time so I thought you'd nailed down the worst of the issues.. I hadn't realized this whole time it was hand-holding [16:45] pitti: thanks for doing a good job on the retracers :) like I said, I hadn't noticed anything in a while.. [16:50] dpm, the pos are in the launchpad export, so it's something in the job that build the langpack sources from the export, do we have logs for that? [16:51] seb128, I can confirm that too, I can see at least gedit.po in the export. langpack-o-matic can probably tell us more, pitti should know if we've got any logs [16:51] right [17:18] $ sudo ifup eth0 [17:18] sudo: unable to resolve host ubuntu [17:18] dhclient: error while loading shared libraries: libirs-export.so.91: cannot stat shared object: Permission denied [17:18] Failed to bring up eth0. [17:18] that make any sense to anyone ? [17:18] stgraber, ^ ? [17:19] hmm, would guess apparmor [17:19] anything in dmesg? [17:21] what's libirs-export.so.91? [17:21] stgraber, good guess. http://paste.ubuntu.com/12604024/ [17:21] now, i guess some background on the possible luser error here. [17:21] i grabed root-lxd.tar.xz for wily [17:22] put it into a ext4 file system. then tried booting it with kernel, initramfs from maas ephemeral [17:22] do you have overlayfs in that mix? [17:22] and [17:22] root=/dev/vda ro console=ttyS0 overlayroot=tmpfs init=/bin/bash [17:22] yes. [17:23] oh, so it's yet another case of apparmor blowing up on overlayfs [17:23] maybe [17:23] but in an odd way [17:23] kinda surprised we've not run into that one with the desktop image [17:23] because most certainly this works in maas booting maas ephemeral images. [17:23] so there might be some luser error in the mix. [17:23] but i dont know wht it woudl be. [17:24] the first error i saw was just that networking-online never fired. [17:24] smoser: so one thing I notice in there is that none of the names contain a leading / [17:24] smoser: which may explain why apparmor can't resolve the paths, though I'm not sure what you may have done on your side to cause that slight difference [17:24] hm. [17:25] I think it's the overlayroot=tmpfs that did it [17:26] nah [17:26] that path works [17:26] but this bug suggests the usual "attach_disconnected" workaround for files not visible in the process's namespace may not be sufficient for overlayfs: https://bugs.launchpad.net/apparmor/+bug/1408106 [17:26] Launchpad bug 1408106 in AppArmor "attach_disconnected not sufficient for overlayfs" [Critical,In progress] [17:26] odd though.. [17:26] it doesnt seem like i'm doing anything that maas doesnt do in its installation of wily [17:27] nah, well sort of its overlayfs use of clone_private_mount [17:27] and that is known working (at least as of a few days ago) [17:27] sarnold: that depends, it is sufficient for regular file accesses, it is not sufficient if the denial is in a mount, or pivot_root [17:28] jjohansen: how about this? < smoser> root=/dev/vda ro console=ttyS0 overlayroot=tmpfs init=/bin/bash [17:29] my current recommendation is to turn apparmor off in any environment where / is on overlayfs, so live media, maas and any relevant installer [17:29] sarnold, yea, i ditched that though, that was only for debugging. [17:29] smoser: ah, sorry for confusing the issue [17:29] yeah. good eyes though [17:29] that's very much of a selinux-sounding answer to the problem, but fact is, this is a problem affecting random paths in a bunch of cases and it'd be really annoying if $RANDOM_SRU starts breaking things because we were depending on something which isn't actually supported [17:30] sarnold: won't fix the apparmor bit, the current work around is using the attach_disconnected flag in the profile [17:31] but if it is failing in a mount (not this case) there is no work around, besides making the affected code unconfined [17:31] * jdstrand notes that we hope to re-engage with overlayfs upstream next month (ie, after current sprint that is in progress) and see what's what [17:32] jjohansen: is there a way to set attach_disconnected for every single profile on the system without having to mangle them all? [17:33] because with / being on overlayfs, it's potentially every single apparmor profile that would run into this and I'm really not sure we want to change them all to set attach_disconnected [17:33] stgraber: no [17:33] so attach_disconnected isn't really an option then [17:35] Hi, would someone mind looking at this patch in the next pilot run? https://bugs.launchpad.net/ubuntu/+source/samba/+bug/1385868 does this perhaps need to be fixed upstream? [17:35] Launchpad bug 1385868 in samba (Ubuntu) "Samba logrotate script uses invalid argument to /etc/init.d/nmdb" [Medium,Confirmed] === athairus is now known as afkthairus [18:01] seb128: it looks to me like the Packages file at http://ddebs.ubuntu.com/dists/wily/main/binary-armhf/ is missing libtrust-store2 [18:01] pitti: it looks to me like the Packages file at http://ddebs.ubuntu.com/dists/wily/main/binary-armhf/ is missing libtrust-store2 - probably among other things [18:02] pitti: wily/main/binary-amd64/ is missing it too [18:21] seb128, pitti: I reported bug 1500557 about it. [18:21] bug 1500557 in ddeb-retriever "Packages file(s)? for wily out of date" [Undecided,New] https://launchpad.net/bugs/1500557 === lutostag_ is now known as lutostag [18:39] @pilot in === udevbot changed the topic of #ubuntu-devel to: Archive: feature freeze | Devel of Ubuntu (not support or app devel) | build failures: http://qa.ubuntuwire.com/ftbfs/ | #ubuntu for support and discussion of precise-vivid | #ubuntu-app-devel for app development on Ubuntu http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://bit.ly/lv8soi | Patch Pilots: cyphermox === Tm_Tr is now known as Guest47262 [19:10] Does Wily image now support 32-bit UEFI on x64 processor? [19:19] bdmurray, thanks for investigating === afkthairus is now known as athairus [20:22] what's the 'correct' way to get the arch-specific part of /usr/lib/${arch}/... [20:25] dpkg-architecture -q DEB_HOST_MULTIARCH is certainly one way [20:25] hmm, the phone doesn't have dpkg-architecture installed... unfortunate [20:32] do we have perl on the phones? that' sjust a 9k perl script.. [20:32] it feels like a useful enuogh thing to include [20:38] sarnold: we have perl-base; we should not have perl; and dpkg-architecture is part of dpkg-dev which has other baggage [20:40] slangasek: oh :/ I didn't think to check where it comes from.. [21:00] slangasek, so any other way to work that out than dpkg-architecture? [21:06] brendand: is coreutils installed? /usr/bin/arch [21:06] TJ-, that's only part of it [21:07] TJ-, in fact i'm not sure that's related to it at all [21:07] on the phone `arch` is "armv7l" [21:08] but i'm looking for something that gives "arm-linux-gnueabihf" [21:08] that comes from the libc build triplet doesn't it? [21:09] TJ-, i guess so [21:14] brendand: no good way to check this at runtime, no, sorry [21:15] brendand: you can indeed have your own local table that maps architectures to paths, but there's nothing that does this mapping for you [21:15] Hi [21:16] slangasek, is there any reason 'armv7l' might not map to 'arm-linux-gnueabihf'? [21:21] Hello, I am new to bug fixing. I would love to fix and patch bugs [21:21] Could you rely on the binutils package? It installs files under /usr/bin/ with the triplet encoded in the filename [21:22] But i am unable to get how to start and where to start? [21:30] bipul: see https://wiki.ubuntu.com/Bugs [21:30] brendand: you might have a 64-bit kernel, and then you don't get 'armv7l' at all but 'aarch64'? so the more reliable check for "what architecture am I on" is "dpkg --print-architecture" [21:37] brendand: the correct way to get the multiarch path is to calculate it at package build time and embed it in your package [21:38] (which may require making your package architecture: any) [21:38] and then you can use dpkg-architecture at build time [21:39] cjwatson, the particular use-case here is to find the path of a binary during a test [21:39] cjwatson, so it will be well after build time [21:39] brendand: I stand by my advice [21:40] brendand: not saying you need to be able to construct the entire path of the binary in question at build time, but you can embed the multiarch bit; or else question why a binary that's needed outside of implementation details of its package is in a multiarch path with no non-multiarch-dependent way to get at it ... [22:04] brendand: Err, "find the binary of a path at runtime"? [22:05] brendand: If that relies on the triplet, you've failed to do something correctly, IMO. [22:05] brendand: Could you elaborate on the problem, rather than the proposed solution? [22:05] s/binary of a path/path of a binary/ [22:06] But I'm not dyslexic... === salem_ is now known as _salem === adrian is now known as alvesadrian