=== salem_ is now known as _salem
=== timchen1` is now known as timchen119
dupingpingplease help me.03:31
=== nudtrobert1 is now known as nudtrobert
=== sunweave1 is now known as sunweaver
=== marcusto_ is now known as marcustomlinson
=== cpaelzer_ is now known as cpaelzer
dholbachgood morning07:28
pittislangasek: yeah, I know; It's ridiculously hard to inspect AMQP queues, much harder than e. g. doing the live logtailing :/07:38
pittislangasek: however, we could publish britney's pending.txt somewhere, which would already be a start07:39
pittiLaney: wow, now that we get mailed about the cloud errors/worker failures it's quite dazzling how many there are..07:42
seb128doko, hey, "libllvm3.7: Add Conflicts/Replaces to libllvm3.7v4 (anticipated backport)." ... what do you mean "anticipated backport"? could that go to Debian as well?09:14
seb128doko, bug #1501300 ... seems like llvm-toolchain-3.4 was built with gcc-4.9 in wily due to an issue with the debian/rules gcc version regexp, do you think fixing that/building with gcc5 is the sort of change that is fine for a SRU?09:33
ubottubug 1501300 in llvm-toolchain-3.4 (Ubuntu) "Wily (15.10) this package got not compiled with __cxx11 support" [Medium,Fix released] https://launchpad.net/bugs/150130009:33
seb128same for bug #151112809:35
ubottubug 1511128 in llvm-toolchain-3.7 (Ubuntu) "Package compiled on wrong compiler for wily" [Medium,Confirmed] https://launchpad.net/bugs/151112809:35
Laneypitti: yeah... what's up with the ssh timeout ones? just the instances failing to come up properly?09:44
=== nudtrobert1 is now known as nudtrobert
pittiLaney: seems they all come from vivid systemd's fsck test09:59
pittiLaney: it timed out, it got handed to another worker, and then killed that as well10:00
pittiLaney: I guess the possibility of chain-killing workers is now a consequence of not giving up on tmpfails10:00
pittibut I'd rather have it this way, this happens much less (systemd/vivid/ppc64el is the only known case to me)10:00
pittiLaney: the others were quota ceiling again10:01
pittiLaney: wrt. http://autopkgtest.ubuntu.com/status/alerts/ -- I retried mpfit (RT to delete the two undeletable nova instances got processed and done)10:02
pittiLaney: systemd/network-manager are these chain-killers, to be investigated10:03
pittiLaney: apw and I found the reason for the linux/ppc64el test dep uninstallability, but there's still a bug somewhere in my code because it re-tries twice and then fails on the already removed apt pin file10:03
pittiso TL;DR: I'm on them, please don't retry those10:03
Laneyokay, thanks for being on top of them!10:04
pittiLaney: I'm now looking into converting these pesky "Temporary failure resolving 'ftpmaster.internal'10:05
pittiLaney: ... to tmpfail, so that they will auto-retry10:05
pittiinstead of having to be retried manually10:05
Laneyis that fatal or can adt itself be retried?10:06
Laneyi.e. do you have to kill the instance when that happens?10:06
pittiLaney: you mean apt?10:06
Laneywell I presume it makes adt-run fail10:06
pittiwe already retry apt-get update10:06
Laneycan you try it again?10:06
pittiwe might try to re-run apt-get install too10:06
pittiLaney: if adt-run exits with a failure, the container is gone, so we have to start over, yes10:07
pitti(but I don't see that as a big problem)10:07
pittiretrying apt-get install itself is a bit more elegant, though10:07
pittiin both cases I simultaneously need to capture apt's output to check for the error message but also have it go to stdout/err for the log10:08
pittiso I might just as well do it locally in the apt-get call10:08
pittiapt-get --quiet update || (sleep 15; apt-get update)) 2>&110:09
pittiwe already do that trick for update, which commonly fail with hash sum mismatches10:09
pittihm, but if the outages last longer than a few seconds, then the test can't do much about it by itself10:13
pittimvo: apt-get always exits with 100 on errors? is there a way to tell apart errors on downloading/fetching stuff from uninstallability or postinst errors?10:15
pittimvo: apart from checking the output, I mean10:16
pittimvo: or maybe some extra "side channel" where it outputs such errors in addition, so that I can leave normal stdout/stderr in place?10:16
pittimvo: like dpkg's --status-fd10:16
* pitti discovers README.progress-reporting in the source and APT::Status-Fd10:19
pittierr, the status-fd messages are translated10:22
apwheh just what you need10:23
pitti-o APT::Status-Fd=2 works10:24
pittibut this doesn't, /tmp/out is empty; sudo apt-get install -o APT::Status-Fd=2 pmount 3>/tmp/out10:24
pittierr, APT::Status-Fd=3 of course10:25
pittiapparently my redirection fu isn't strong enough10:25
apwpitti, some LANG=C foo tooo to make sure its parable10:25
pittioh, sudo, hang on10:25
pittiyep, works without sudo and in a root shell, sorry for the noise; so indeed weak redir fu :)10:26
pittimvo: so, found it, unping10:28
pittiso I think if the file contains ^dlstatus: but does not contain ^pmstatus:, I can assume downlaod failed and it's worth trying again10:30
pittiwith sleeping 10 s in between, and give up with tmpfail after the third failure10:31
arahello all,10:41
araI have a general question about contributing to Ubuntu through bzr branches10:41
mvopitti: hi, what exactly is the problem you need to solve? I guess I can look into giving you different exit codes for certain failures10:42
araCurrently, there is no xenial branch for partman efi: https://code.launchpad.net/ubuntu/+source/partman-efi10:42
aracan we expect that branch to exist at some point? how/when that happens?10:42
pittimvo: we have a lot of failures like https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-xenial/xenial/armhf/x/xen-tools/20151111_234936@/log.gz10:42
pittimvo: so if a package fails to install due to postinst bugs, bad dependencies etc., I want the test to fail10:43
pittimvo: but if it fails to download, or these DNS errors, I want to retry or make it a temporary testbed failure insted of a test failure10:43
pittimvo: but APT::Status-Fd=3 actually works well enough10:43
* pitti checks if that's available in precise10:44
mvopitti: ok, so a different exit code for download failures? that sounds not too terrible, let me check10:44
pittimvo: well, I need something that works on all releases; so it might be nice for the future as these retry loops become much simpler10:44
mvopitti: aha, ok. if status-fd is good enough, sure. it should be in precise, its availalbe since a long time (~2007ish I would say)10:44
pittiyep, works in precise10:45
pittiso, that's good, thanks!10:45
pittimvo: so I guess I'll use "apt-get install -o APT::Status-Fd=3 ... 3>&2 2>&1" and capture stderr10:49
caribouhmm, looks like my upgrade to Xenial was not a success : lost sound, external display, USB keyboard & I don't get a prompt upon boot for my encryption password10:52
arahey dholbach, can I ask you a question related to contributing to Ubuntu through bzr branches?10:53
pitticaribou: wow -- was it interrupted in the middle or so?10:53
pitticaribou: the worst that I can say is that gedit now looks strange, so it's not generally broken for everyone10:53
dholbachara, sure10:54
cariboupitti: no nothing to report10:54
aradholbach, Currently, there is no xenial branch for partman efi: https://code.launchpad.net/ubuntu/+source/partman-efi10:55
aracan we expect that branch to exist at some point? how/when that happens?10:55
cariboupitti: I'm going to reboot with  a clean syslog & I'll be back10:55
dholbachara, no idea - let me ask somebody10:57
dholbachpitti, cjwatson (or anyone else): do you know why there's no xenial branch for https://code.launchpad.net/ubuntu/+source/partman-efi yet? (ara just asked.)10:57
pittiI don't know, sorry; I don't know whether anybody still maintains the UDD branches in LP, but it seemed pretty dead to me10:58
pittito be blunt, they've been a mis-design right from the start, and a lot of imports tend to be behind; I thought there was a new plan to have something like dgit10:58
dholbachpitti, so you'd recommend to just send a patch?10:59
pittidholbach: personally, yes (I know other sponsors actually like the UDD branches, so don't consider that a "project opinion")10:59
pittipersonally, a debdiff to a sponsoring bug is ten times faster and simpler than wrestling with the UDD patch maze10:59
pittiand no sponsor I know of dares to actually push these branches11:00
dholbachara, ^ does that help?11:00
arayes, it does11:00
aradholbach, the packaging guide still says that UDD is the way to go :)11:00
aradholbach, pitti: thanks!11:01
pittiara: the other option is to work on the Debian git, but we have a delta, so this mostly just makes sense for more convenient development if you want to use git diff, stashing, etc.11:01
pittiand then attach the git format-patch to the bug11:01
dholbachara, I guess it needs a bit of a broader discussion - but yeah, I agree11:03
Laneydholbach: have the candidates for CC written any kind of platform?11:07
LaneyI don't know how to choose between most of them tbh11:07
cjwatsondholbach: right, more or less as pitti says.  We haven't brought them up yet, and conceivably, we might be close enough to being able to have dgit working that it might not be worth the effort.  And the auto-import branches have been so badly broken in so many ways for so long ...11:09
dholbachsome might have something up on their website, but as we were a bit late already, I didn't want to go through another round of "please write something and send a link back" before we announce11:09
dholbachcjwatson, cool - thanks!11:09
cjwatsonara: I have been complaining about that in the packaging guide for years :P11:09
aracjwatson, :)11:10
cjwatsonnothing ever happens because rah rah revision control (even if it's horribly broken)11:10
dholbachok... let's have a discussion about it and change it11:11
dholbachI'm not sure if I'm seen as blocking this... because I'm happy for it to be changed11:11
cariboupitti: systemd complains about some apparmor issue : http://paste.ubuntu.com/13237626/11:13
=== _salem is now known as salem_
pitticaribou: ah, I get that too11:33
pittisudo systemctl status apparmor -l11:33
pittiNov 12 08:34:00 donald apparmor[476]: AppArmor-Analysefehler f?r /etc/apparmor.d/usr.bin.webbrowser-app in /etc/apparmor.d/usr.bin.webbrowser-app in Zeile 26: >>/usr/share/apparmor/hardware/graphics.d<< konnte nicht ge?ffnet werden11:33
pittiso apparently an error in webbrowser-app's apparmor profile11:33
pittibut that should be moderately harmless11:33
pittiwhy do I have webbrowser-app installed in the first place..11:33
seb128pitti, we install it by default (should be cleaned out but we noticed that just before xenial and it was too late for such changes)11:40
pittiseb128: s/xenial/wily/ I  suppose11:41
seb128it's used by webapps but we don't install any by default so we don't need it on the iso11:42
cariboupitti: I found my problem11:44
cariboupitti: linux-image-extra for .18 is missing in the Xenial archive11:44
cariboupitti: apw is looking into it11:44
pitticaribou: oh indeed; I still have -16 installed11:45
cariboupitti: I also had  .17 so going back to that one fixes my usb/sound/external display problem11:45
pitticaribou: .17 is in -proposed still11:45
pittinever ever (and I mean EVER) enable devel-proposed on a real machine11:46
pittidevel-proposed by definition are all packages which are broken11:46
cariboupitti: I don't have it enabled.11:46
cariboupitti: apparently it is a miss from a CVE upload11:46
pitticaribou: oh, wily now has a newer kernel than xenial? nice..11:46
cariboupitti: no, xenial has linux-image- but it is missing linux-image-EXTRA-
pitticaribou: ok, so missing -extra does explain missing sound, graphics, and stuff11:47
cariboupitti: yes11:47
cariboupitti: thanks for diwic for nailing it down11:48
caribouok, rebooting to see if I get my password prompt11:48
pittihm, so wily-updates has 4.2.0-18.22, xenial has 4.2.0-16.19 , xenial-proposed has 4.2.0-17.2111:48
pitticaribou: that explains why I didn't see that, as I dist-upgraded to xenial right away, before I got wily-updates bits11:49
pittidiwic, caribou: thanks for tracking it down!11:49
cariboupitti: well, that's the reason for running/upgrading to Xenial isn't it ?11:49
pitticaribou: well, usually I'd expect the breakage to be in the devel series, not in the stable series11:50
pittiin devel it would be "meh", but devel is fine11:50
pittiit seems we completely screwed wily with that11:50
pittiand that's in the "disaster" category11:51
pitticaribou: or did I misunderstand this somehow? did you have -18 extras in wily, but it somehow got lost on the xenial update?11:51
pittilinux-image-extra-4.2.0-18-generic | 4.2.0-18.22 | wily-security | amd64, i386, ppc64el11:52
pittiah no, *phew*11:52
pitticaribou: so your -18 kernel got partially removed during the upgrade then?11:53
apwpitti, the tl;dr was that the install of -18 was manual and not done completly, so all is understod and we're fine11:53
pittiapw: ok, good11:53
apwpitti, he wanted -18, but was on xenail where it did not exist11:53
cariboupitti: no, dist-upgrade was complaining it couldn't find it in the archive so I went back to wily and manually installed _only_ the linux-image11:53
apwpitti, and i've requested that be copied up to close this hole11:54
cariboupitti: then dist-ugprade11:54
pittiLaney, mvo: this works nicely: http://anonscm.debian.org/cgit/autopkgtest/autopkgtest.git/commit/?id=609c9912:07
pittiLaney: and the .bomb() instead of .badpkg() will make it a tmpfail instead of a test fail, so hopefully these shoudl be gone12:07
* pitti runs a mass-retry to mop up the current failures12:08
pittiI'm sure that we now find some tests which are genuinely broken but trigger this :)12:08
mvopitti: heh :)12:09
seb128Laney, did you get a reply from dholbach earlier about the community council eventual programs? (didn't see you but there were cross discussions)12:37
Laneyseb128: "eventual programs"? nothing further no12:38
Laneybut I don't know what one of those is. :P12:38
seb128sorry, I guess franglish12:38
Laneyplatform is the thing I asked for12:38
seb128you asked if candidates eventually had a platform12:39
Laneydon't think there is any central list of them12:39
seb128we use "programme" for "platform" in french12:39
seb128I though "program" would work in this context in english12:39
seb128sorry ;-)12:39
Laneyno worries12:40
seb128like "having an agenda"12:40
cjwatsonIt kind of works but wouldn't be the usual term12:41
seb128I see12:42
dholbachseb128, Laney: sorry - maybe I should have been clearer earlier.... we were a bit late with the election and I wanted to get it announced as opposed to mailing everyone again, asking for a platform to be written, pages updated, etc12:42
seb128dholbach, no worry, but does it mean it's likely that some people will write something?12:43
seb128or said differently is waiting likely to provide more details to make an informed choice?12:43
dholbachseb128, I haven't pinged people yet, but I could do that - and send a mail with links to their wiki page in an update?12:44
seb128dholbach, that would be nice12:44
=== nudtrobert1 is now known as nudtrobert
seb128dholbach, don't feel like you have to do it, I was just asking in case because I don't know some of the candidates, but I opened their launchpad page which already gives some kind of idea12:47
dholbachseb128, I would... but the wiki doesn't let me :-P12:57
Laneydoko: you here? can you take a look at phonon-backend-vlc maybe?13:47
Laney(build failure, -fPIC stuff, I can't get it to build)13:48
jdstrandpitti: the webbroswer-app bug is known. oSoMoN is working on it13:50
jdstrandbug #151143913:51
ubottubug 1511439 in webbrowser-app (Ubuntu) "webbrowser-app apparmor policy fails to load on desktop" [High,In progress] https://launchpad.net/bugs/151143913:52
pittijdstrand: thanks13:56
jdstrandpitti: do you need more for migration or is that enough?13:57
jdstrands/more/more info/13:57
pittijdstrand: oh, I was looking at that in the context of caribou's upgrade b0rkage, not proposed-migration13:57
jdstrandpitti: ah. well, in either case, you can install apparmor-easyprof-ubuntu on the system or disable the profile with aa-disable14:00
jdstrandand that should fix things. oSoMoN's fix will (obviously) also fix it, but it hasn't landed yet14:01
pittijdstrand: TBH I just purged webbrowser-app; seb128 said it's going away from the default install anyway, and I have no need for it14:01
jdstrandpitti: heh, that is another option :)14:01
jdstrandpitti: ok, looking at proposed migration, the regression has nothing to do with apparmor: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-xenial/xenial/amd64/s/systemd/20151112_015036@/log.gz14:03
jdstrandFAIL: test_hotplug_dhcp_ip6 (__main__.TestNetworkd)14:03
pittijdstrand: no, that's a race condition in the networkd test, I'm aware of it14:03
jdstrandAssertionError: nameserver not found in /etc/resolv.conf14:03
pitti(either in the test, or in the resolvconf integration, not sure yet)14:03
jdstrandpitti: does that mean you'll handle the proposed migration for apparmor then?14:04
pittijdstrand: yes, will do14:04
jdstrandpitti: thanks! :)14:04
oSoMoNjdstrand, pitti: my fix for webbrowser-app is awaiting QA verification, will hopefully land soon14:06
pittidoko: asking you because TIL: puppetmaster-passenger started to become uninstallable because it depends on the removed libapache2-mod-passenger; do you have any idea about that, or is that server team matter?14:12
pittiah, this was fixed in Debian14:13
pittiso, merge o'clock14:13
pittidoko: I can do the merge if you don't care, but would like to ask you first14:14
yofelLaney: do you plan to merge lintian anytime soon?14:25
yofelwith debhelper updated all libraries in the kubuntu ci throw lintian errors about missing ldconfig calls. So if you don't plan to merge it soon I would give it a try.14:25
popeyAnyone else had apport just sit there chewing a cpu for _ages_ on xenial?14:35
popey  PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND14:35
popey 5386 root      20   0  164460  82176   8768 R 100.0  0.5 124:53.76 apport14:35
popeyahh, it's dealing with a caja crash, will poke flexiondotorg :)14:36
pittipopey: is it actually writing a .crash file in /var/crash and is the crashed process still existing (in 'D' state)? or is it trapped in a post-crash loop?14:36
popeynothing in /var/crash14:36
popeywas just sat there in a loop14:36
popeyalan     27679  0.0  0.3 1211948 62316 ?       Sl   Nov10   0:07 caja14:37
popeythats the process it was dealing with14:37
pittipopey: can you show me the tail of /var/log/apport.log? This caja process didn't even crash14:38
popey-rw-r----- 1 root adm 0 Nov 11 07:39 /var/log/apport.log14:39
popeyits a days old process, maybe an older apport log...14:39
ogra_definitely a short tail14:39
pittipopey: what's the full command line of that apport process?14:40
popeyroot      5386 99.8  0.5 164740 85876 ?        R    12:29 125:47 /usr/bin/python3 /usr/share/apport/apport 27679 11 0 2767914:40
popeyhttp://paste.ubuntu.com/13238745/ is the previous apport log14:40
pitti"another apport instance is already running, aborting", ah14:40
pittithis makes no sense at all14:41
pittiso caja crashed with SEGV (signal 11), the kernel called apport, but the process is still alive; it should go into 'D' and then die after core dumping14:41
pittipopey: can you attach strace to apport for a bit and see what it's looping on?14:42
popeyI'd already killed it, sorry :(14:42
* pitti wonders if it's continuously reading some junk from the kernel as core dump data14:42
popeyIt may happen again, I have seen this before14:42
popeyand if it does, I'll trace it14:42
Laneyyofel: I'll get to all my merges soon, been a bit busy shepherding some transitions15:06
yofelLaney: ok, then I'll leave it to you as I would need a sponsor for it.15:12
=== adrian is now known as alvesadrian
rbasaktdaitx: how is the squid3 merge going?15:42
rbasaktdaitx: no hurry, just checking it's not stuck.15:42
tdaitxrbasak, hi there! well, it is stuck, but just because we didn't have time to work on it during the sprint =)15:43
rbasaktdaitx: OK, no problem.15:43
dokopitti, puppet is now demoted16:00
pittidoko: i. e. does that translate to "I don't care, go ahead and merge", or something else? :-)16:12
rbasakdoko: I wasn't aware of any plan around puppet. Is that intended to be permanent?16:14
dokopitti, the former ;p16:18
pittidoko: ok :) I'll do that then to fix that again16:19
rbasakdoko: I wasn't aware of any plan around puppet. Is that intended to be permanent?16:24
smoserdoko, can you give some background on that ? i think that i've seen something around it, but dont remember the details.16:25
infinityWhatever was pulling it into main must have stopped...16:25
infinitydoko: Yeah, why was that done? :P16:26
infinitydoko: supported != shipped, so that has no effect on images.16:27
smoserhey. i need to fix http://pad.lv/151566116:29
ubottuLaunchpad bug 1515661 in cloud-initramfs-tools (Ubuntu Precise) "SRU:add cloud-initramfs-copymods to 12.04" [Medium,Confirmed]16:29
smoserthe easiest thing to do would be to take the trusty-updates version and get it to precise.16:30
smosertrusty-updates is 0.25ubuntu1.14.04.116:30
dokorbasak, infinity: puppet was only needed by IS, and isn't anymore. I didn't get any reply from the server team that it is still needed.16:30
rbasakdoko: I wasn't aware that you asked. Where did you ask?16:30
rbasakdoko: I'd prefer us to make a considered decision on whether we want it in main rather than have it ripped out from under us.16:30
dokorbasak, nothing is permanent until the release is done16:33
rbasakdoko: I wasn't aware that you asked. Where did you ask?16:33
dokorbasak, by email16:33
rbasakTo whom?16:33
seb128doko, hey, did you see my llvm/SRU question this morning?16:36
dokoseb128, yes, but I didn't look at it yet. the issue is that backported llvm 3.7 packages need a different library package name when built with GCC 4.9 or earlier16:38
seb128doko, speaking of gcc, apparently the wily packages are built with 4.9 instead of 5 due to an issue with the debian/rules regexp issue, unsure if changing the gcc version is material for a SRU?16:39
tdaitxslangasek, when configuring my system to work on a high dpi display I noticed that console-setup, console-setup-linux, and keyboard-configuration have upstart scripts, but only keyboard-configuration has a systemd service, thus the font I selected on console-setup is never applied on boot time (there are open bugs for that)16:39
tdaitxis that the expected behavior or somtehing that needs fixing?16:39
dokoseb128, afaics, it's not yet used, so maybe nobody cares (and somebody should address the ftbfs on i386 and powerpc)16:40
seb128doko, there is a bug reported so I guess some people care16:41
seb128and ack on the build issues16:41
tdaitxslangasek, apart from that, I find it a bit weird that keyboard-configuration init/systemd service is named console-setup and its initram hook also handles font and depends on stuff setup by setupcon (setupcon is from console-setup, which actually depends on keyboard-configuration, not the other way around)16:47
slangasekMirv: do you have any clues about this x86-only build failure in phonon-backend-vlc? it looks like something internal to cmake+qt5 rather than a bug in phonon-backend-vlc's code:16:49
slangasekIn file included from /usr/include/x86_64-linux-gnu/qt5/QtCore/QtGlobal:1:0,16:49
slangasek                 from /«PKGBUILDDIR»/obj-qt5/CMakeTmp/check_qt_visibility.cpp:1:16:49
slangasek/usr/include/x86_64-linux-gnu/qt5/QtCore/qglobal.h:1052:4: error: #error "You must build your code with position independent code if Qt was built with -reduce-relocations. " "Compile your code with -fPIC (-fPIE is not enough)."16:49
slangasek #  error "You must build your code with position independent code if Qt was built with -reduce-relocations. "\16:49
slangasektdaitx: console-setup/keyboard-configuration have a very long history, and it's possible the current state includes some historical reasons.  And if the configured console-setup font is not applied at boot, yes, that's a bug and quite possibly related to this16:49
slangasektdaitx: cyphermox would have some recent context about console-setup; pitti might know a thing or two about how systemd expects to set console fonts16:50
dokoslangasek, this popped up a lot during the last cycle. iirc, the "solution" was to remove -fPIE and add -fPIC.16:50
dokosbeattie, ^^^16:51
slangasekdoko: well, but this package isn't passing -fPIE16:51
dokobut -fPIC?16:52
slangasekdoko: no - I mean that -fPIE is being passed but not by this package16:52
slangasekmaybe it comes from /usr/share/phonon4qt5/buildsystem/FindPhononInternal.cmake16:52
Odd_BlokeIs there an email from gaughen_ sitting in the ubuntu-devel@ moderation queue?16:54
slangasekthere's a KDE4_ENABLE_FPIE16:54
Odd_Bloke(And if so, who can unblock it?)16:54
slangasekOdd_Bloke: I think cjwatson has moderation privs there; not sure who else might16:55
Laneyme, nothing there16:55
LaneyOdd_Bloke: ^16:57
gaughen_Laney, sweet, thanks!16:58
nemoFWIW, I finally solved the problem that I came by here to figure out:16:59
ubottuLaunchpad bug 1268770 in vmware-view-client (Ubuntu) "Error loading shared library for smart card authentication to server" [Undecided,Confirmed]16:59
nemoall better16:59
nemo'course, IMO there's a couple of broken packages there, but whatev ☺16:59
nemoworks for me! 😝16:59
doko$ simple-scan17:02
dokosimple-scan: error while loading shared libraries: libpackagekit-glib2.so.16: cannot open shared object file: No such file or directory17:02
dokoseb128, ^^^ can you reproduce this? or was this something part of wily-proposed which didn't go into the release?17:03
seb128$ dpkg -S libpackagekit-glib2.so.1617:04
seb128libpackagekit-glib2-16:i386: /usr/lib/i386-linux-gnu/libpackagekit-glib2.so.1617:04
seb128do you have that package?17:04
doko$ dpkg -S libpackagekit-glib2.so.1817:05
dokolibpackagekit-glib2-16:amd64: /usr/lib/x86_64-linux-gnu/libpackagekit-glib2.so.1817:05
dokolibpackagekit-glib2-16:amd64: /usr/lib/x86_64-linux-gnu/libpackagekit-glib2.so.18.0.017:05
dokoso, maybe left over from using -proposed during the GCC 5 transition17:05
seb128doko, yes, that version of packagekit was in proposed only and breaking aptdaemon and never migrated/was removed17:07
dokoahh, I remember ...17:07
seb128but the binary should depends on the lib it uses17:07
slangasekhow does anyone debug cmake, ever?17:07
seb128so looks like there is a bug17:07
Laneyslangasek: I just pinged mitya57 about this and he said he'd look, FYi17:08
Laneyas a more qualified cmaker than me17:08
slangasekLaney: alright; my question stands17:08
slangasekthis is untraceable spaghetti17:08
slangasekcheck_cxx_compiler_flag(-fPIE HAVE_FPIE_SUPPORT)17:09
slangasekso OBVIOUSLY that causes all tests to immediately start using that flag17:09
mitya57I don't understand why this bug doesn't happen in Debian17:10
mitya57In simple case one just needs to add something like set(CMAKE_CXX_FLAGS "-fPIC") and that will work. But looks like it's a bit more complicate here.17:10
* mitya57 is not a CMake expert unfortunately17:11
slangasekI can't find anything in the cmake code that explains why -fPIE is being used17:11
slangasekmaybe you'll have better luck than me17:11
slangasekmitya57: Debian version of libphonon4qt5-dev doesn't include the check for -fPIE17:14
mitya57ah, I thought that warning was coming from Qt itself17:15
mitya57Actually I believe adding actually set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} ${Qt5Widgets_EXECUTABLE_COMPILE_FLAGS}") to CMakeLists.txt will just work17:15
slangasekmaybe it's worth checking if this is fixed by a phonon merge from unstable17:17
cyphermoxtdaitx: what about console-setup/keyboard-configuration?17:17
slangaseksince there are (surprising) differences in the installed cmake rules, and Debian doesn't have this problem17:17
tdaitxcyphermox, well, I set up a new font and keyboard layout for the new computer, but only the keyboard layout is being applied on boot... I'm investigating what I might be missing in my initramfs config, but anyway I noticed that console-setup, console-setup-linux, and keyboard-configuration have upstart scripts, while only keyboard-configuration has a systemd service, so the console font is not being set17:20
* Laney stares at gstreamer-vaapi17:22
Laneyah right, probably just skew17:23
=== sarnold_ is now known as sarnold
cyphermoxtdaitx: interesting. text-mode console font?17:30
tdaitxcyphermox, yes17:31
cyphermoxas you expect this would be applied in initramfs, before systemd, no?17:31
cyphermoxwhich font, for fun? if I want to install it here?17:31
tdaitxit should be17:32
tdaitxcyphermox, ^17:33
cyphermoxyep, will apply this here on my temporary laptop, while the tech finishes changing the mobo17:33
tdaitxcyphermox, you might want to use a smaller one =)17:35
tdaitxin case you dont have a high dpi display17:35
seb128tjaalton, is anyone looking at bug #151097017:42
ubottubug 1510970 in xserver-xorg-video-intel (Ubuntu) "Intel driver crashes on Ubuntu 15.10" [Medium,Confirmed] https://launchpad.net/bugs/151097017:42
seb128seems lamont is able to reproduce easily17:42
lamontwell, easily is more one of "I do things, and it gets mad sometime with in about 10 minutes of normal work on my laptop"17:43
lamontbut yeah, it reproduces on an alarmingly easy basis17:43
cyphermoxtdaitx: I see so well that this is a great font size for me ;)17:43
cyphermoxyay, I gotz new variables now.17:44
tdaitxcyphermox, sorry, my battery ran out17:44
tdaitxcyphermox, heh, nice17:44
cyphermoxtdaitx: so, I agree, needs some systemd magic17:44
cyphermoxbut it should also get applied before then in initramfs and it doesn't seem to be17:44
lamontdbg package installed, laptop rebooting for a known-fresh start, let see how hard it falls over now17:44
tdaitxcyphermox, indeed, I'm trying to figure out why17:45
cyphermoxit should be run setfont in init-top though17:45
bdmurrayhallyn: The T upload for bug 1511830 seems to be missing.17:47
ubottubug 1511830 in libvirt (Ubuntu Trusty) "apparmor denies VM startup when image is network mounted" [High,New] https://launchpad.net/bugs/151183017:47
hallynbdmurray: correct17:53
hallynwhen i uploaded those on 11/04 i was still waiting for libvirt to clear t-proposed17:53
hallynin fact i still am17:53
Unit193hallyn: Re: ubuntu-vm-builder.  Looked into vmdebootstrap?17:53
hallynUnit193: in what sense?  I encourage anyone rdep'ing on vmbuilder to switch over :)17:56
hallyni don't think it was around last time i tried to drop vmbuilder17:57
hallyn(from the archive)17:57
hallyncurrently rdeps are sandbox-upgrader and auto-upgrade-tester17:57
Unit193Aha, so yes.  Nice.17:59
hallynany chance you'd have time to work on proposed debdiffs to move those over? :-)17:59
slangasekLaney, mitya57: well hey, it seems that this same failure causes phonon itself to FTBFS; so the problem lies deeper yet17:59
LaneyI'm sad that I don't get to do the last upload. :)18:00
* Laney has to go18:00
Unit193hallyn: That's not actually directed at me is it?18:00
Unit193(That is, never touched those, not a core dev.)18:05
lamontseb128: so much for "trivial reproduction"18:06
lamontonce I get it to die, I'll update 151097018:06
seb128lamont, :-(18:06
lamontseb128: it did seem to always be around a popup window being created, fwiw18:07
seb128write that on the bug when you get the bt18:07
seb128then we can do some nagging upstream ;-)18:08
lamonttbf, if -dbg makes it go away, ... :D18:10
seb128I doubt it, -dbg doesn't change runtine18:10
seb128those are just files on the side that gdb uses18:10
* lamont otp meeting18:13
hallynUnit193: s'ok, you can still post debdiffs :)18:20
lamontseb128: tbf, when I installed -dbg, it upgraded the driver too18:22
lamontI'm tempted do downrev and grab the matching -dbg18:23
seb128lamont, it's likely that the update fixed it, dholbach had a similar issue and tjaalton said the new git snapshot in xenial included fixes for some similar issues18:24
seb128we should still find the fix for that issue and SRU it18:24
seb128so having a debug bt of the old version would be useful18:24
seb128could help to find the corresponding upstream bug/commit18:25
lamontreboot for cleanstart with old version18:30
tdaitxcyphermox, ok, there is a mismatch between the console-setup (re)configure script and the initramfs hook+init-top18:33
tdaitxcyphermox, for a start, the configure script setups the size as 16x32 and writes that to /etc/default/console-setup as FONTSIZE. AFAICS the font name is in the format 32x1618:34
tdaitxsetupcon does detect and handles that, inverting it to 32x16 before searching for the font file, then copies that file to /etc/console-setup18:35
tdaitxthe initramfs hook and init-top scripts for console-setup use the value from /etc/default/console-setup (16x32) so they wouldn't find the font18:36
tdaitxand, second, the font is saved to /etc/console-setup as a .gz file, but the initramfs hook and script expect a .psf file not a .gz file18:41
tdaitxcyphermox, also, /etc/default/console-setup has an example for using FONT='font1 font2', that works fine for /bin/setupcon but it does not work as expected for both console-setup initramfs hook and init-top script18:47
tdaitxseems like console-setup needs some love18:47
tsimonq2is there a channel in which to ask about the Launchpad API? #launchpad maybe?18:48
cjwatsontsimonq2: #launchpad is fine18:48
tsimonq2cjwatson: ok, thanks18:49
NoskcajWhen are the UDD branches expected to appear?18:56
cjwatsonNoskcaj: it is possible that they will not.  we're still deciding18:57
cjwatsonit is a fair bit of work to bring them up and it's not clear it's worth it given how badly broken they are in so many cases18:58
cjwatsonit depends somewhat on whether we can make progress with dgit in a reasonable time18:58
Unit193What about something ala sources.d.net?  And dgit, hrm..18:59
cjwatsonUnit193: I would love to do sources.d.n but it's not clear when :-)19:00
cjwatsonI think dgit is more urgent19:01
Unit193cjwatson: ..And all those other things you need to fix first, as well as all those other things you'd like to do.  As of yet I've ignored dgit because I don't have upload perms, is there a testing ground for it?19:01
cjwatsonUnit193: not quite that far along yet, current state is https://code.launchpad.net/~cjwatson/launchpad/sourcefileurls-include-meta/+merge/27692819:06
Unit193(I was referring to Debian's use in terms of ignoring it.)19:06
Unit193Ah, cool19:06
cjwatsonthen I need to test out the WIP patch I have to make dgit support Launchpad at all (it needs to query its API for a few things, mostly), and then we have two things to do roughly in parallel, which are (a) figure out how on earth to implement something akin to dgit-repos-server's push policy in turnip (the LP git backend), and (b) make lp:ubuntu/+source/<source> git repositories work, mainly hooking them up to upload permissions19:07
cjwatsonand then figure out and deploy a mass importers19:08
cjwatsonit's not intractably complex but there's a reasonable amount to do19:08
tjaaltonseb128, lamont: oh right, apparently the new snapshot works better, so the fix should ne bisectable19:15
Mirvslangasek: Laney mitya57: it should be patched to ban -fPIE like required with GCC5 (on x86, where Qt autodetecs and enables the -reduce-relocations option qtbase). something similar to what you did Laney here https://launchpad.net/ubuntu/+source/poppler/0.33.0-0ubuntu319:20
slangasekMirv: do you know where this -fPIE is coming from in the first place?  This is somewhere deep in the build-dependencies; we shouldn't patch every leaf package that's getting bitten by wrong cmake checks19:21
Mirvanother example http://launchpadlibrarian.net/214075540/gnuplot5_5.0.1%2Bdfsg1-2build1_5.0.1%2Bdfsg1-2ubuntu1.diff.gz19:22
Mirvslangasek: I'm not sure, those two examples were from hardening flags, +all includes -fPIE19:23
Mirvmost leaf packages, also CMake, have worked without changes19:23
Mirvand in this example, which may be the last one that was done during wily, set -fPIE itself http://launchpadlibrarian.net/213155371/pay-service_2.0.0%2B15.10.20150727-0ubuntu1_2.0.0%2B15.10.20150730-0ubuntu1.diff.gz19:24
slangasekit's not coming from either phonon or phonon-backend-vlc; both packages FTBFS but the -fPIE is coming from somewhere else19:24
Mirvslangasek: phonon has  check_cxx_compiler_flag(-fPIE HAVE_FPIE_SUPPORT) and that would look to be there in the phonon-backend-vlc build log19:30
slangasekMirv: yes, there's a check for whether the compiler flag is supported; but that shouldn't cause it to be used automatically for all tests...19:31
Mirvif it has it, it seems to set (KDE4_CXX_FPIE_FLAGS "-fPIE")19:31
Mirv(reading FindPhononInternal.cmake)19:32
Mirvbut there's also KDE4_ENABLE_FPIE coming from somewhere19:32
davmor2Mirv: what are you even doing online isn;t like 2am for you? ;)19:33
slangasekMirv: but doing set(KDE4_ENABLE_FPIE) in phonon-backend-vlc's CMakeLists.txt doesn't seem to do any good; and I can't find anything that's setting it, anywhere under /usr19:33
slangasekthere's basically *nothing* that says PIE, in the build tree or in the build-deps19:34
Mirvdavmor2: going to sleep :) not 2am, just early mornings19:35
elbruspitti: I missed that encoding error, let me think about that one (but it works on other archs, so I think it is not correct to fail on it.19:46
Mirvslangasek: I need to go, but this sounds related https://git.reviewboard.kde.org/r/123874/diff/2#index_header19:48
Mirvalso in Debian http://anonscm.debian.org/cgit/pkg-kde/kde-req/phonon.git/commit/?id=7b0e6608211707c91b52833adc23f6d9295a691619:51
Mirvmaybe this "small" commit too http://anonscm.debian.org/cgit/pkg-kde/kde-req/phonon.git/commit/?id=c78af618d090b3ee7fd9e124c25e1568b20fa9b319:52
NoskcajAre we ok to sync gpgme1.0 from debian? the default install now carries pinentry-gnome3 and gnupg220:04
slangasekMirv: aha!  strange that I didn't see that in the diff with Debian...20:11
rodrigcis anyone here working on Python 3 ( https://wiki.ubuntu.com/Python/3 )?20:28
=== salem_ is now known as _salem
tewardwhen someone files a bug, is there a reason apport wouldn't work with the package hooks21:04
lamonttjaalton: seb128: sadly, I left the laptop power supply at home this morning, so ... :(21:35
=== alai888 is now known as alai
=== Elimin8r is now known as Elimin8er

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!