[04:42] Good morning [04:42] good morning pitti :) sure sign that it's time to be done working, hehe [04:43] that was a nice networking email, certainly a lot to take in in one go.. [04:45] hey sarnold [04:45] sarnold: the netplan stuff? [04:58] infinity: would you mind marking https://launchpad.net/ubuntu/wily as obsolete? [04:58] cjwatson: ^ I think that's the reason for the ddeb-retriever spam [06:16] pitti: Not yet. [06:16] pitti: I'm doing some removals first. [06:40] pitti: I'll sort it out in the morning. [06:40] infinity: thanks [06:55] pitti, morning, I know I'm becoming a bore, but could you please have a look at autopkgtest results here https://requests.ci-train.ubuntu.com/#/ticket/1575 - it seems some of the runs have disappeared - the two regressions in yakkety need a restart with all-proposed [07:01] Saviq: done [07:02] pitti, thank you, glad it doesn't take you too long - any idea why the disappearing runs? [07:02] Saviq: might still be bug 1588566 -- this keeps getting pushed down my todo list, sorry [07:02] bug 1588566 in Auto Package Testing "autopkgtest results go missing for a long time after the test completes" [Undecided,In progress] https://launchpad.net/bugs/1588566 [07:03] ack [07:13] did the installer create a separate /boot at some point in the history? [07:14] by default [07:14] with some scenarios, like lvm or encrypted root [07:15] or presumably zfs now, as AFAIK you cannot boot from that yet [07:15] seems to work fine :) [07:15] anyway, trying to "argue" with a friend who's pissed because /boot is "always" too small [07:15] when upgrading [07:16] because we do a bad job at removing old kernels? [07:16] yeah [07:16] my wife's stock 14.04 install has the same problem, for some reason the auto-removal doesn't work [07:17] so no size of /boot will ever be large enough for that, the root cause is our kernel inflation [07:17] right [07:17] pitti: yeah, netplan :) very ambitious :) [07:17] this is a desktop machine, probably installed with the alternate-installer at some point in the history === hikiko is now known as hikiko|afk === sinzui_ is now known as sinzui [08:38] yeah, the installer gives the impression that lvm is more flexible, when in fact it causes issues like this :) [08:41] xnox, please could you merge netcfg (and fix the GCC 6 build issue)? [08:55] coreycb, please could you fix the python-taskflow ftbfs? === JanC is now known as Guest77763 === JanC_ is now known as JanC [10:18] pitti, what do we do with missing s390x deps https://requests.ci-train.ubuntu.com/static/britney/ticket-1575/landing-073-yakkety/excuses.html ? Mirv mentioned you discussed this with robru yesterday? === hikiko|afk is now known as hikiko [10:49] Saviq, I thought it should be available, no? [10:50] xnox, it's the upstart removal fallout I believe [10:51] oh, ok. yeah, cause package is available. [10:51] doko, ack [11:05] Saviq: I suppose we need to remove ubuntu-system-settings/s390x? [11:07] Saviq: err, there are no ubuntu-system-settings s390x binaries in Ubuntu [11:07] I suppose the PPA builds it for some reason [11:07] pitti, we might need to, if there's no other way, but that will be a gift that keeps on givin'... don't we have a plan to break this dependency chain somewhere? [11:07] pitti, probably because it doesn't have a build-dep on anything that's not on s390x any more... [11:08] hi! [11:09] I would like to report an apparent mismatch between some outputs in w, who and uptime commands [11:09] this is my situation: http://paste.ubuntu.com/21877593/ [11:09] Saviq: but it also was not built in https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-073/+packages on s390s [11:09] Saviq: I think yesterday's problem was that there was some stale s390x binary in some PPA; robru cleaned it up for a different case [11:09] both w and uptime indicate 5 users; but w lists just 2 users [11:09] and who does the same [11:09] pitti, looks built https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-073/+sourcepub/6764432/+listing-archive-extra ? [11:10] not sure what exactly he did [11:10] (my system is Xubuntu 14.04) [11:10] (the commands give the same results when launched with or without sudo) [11:10] Saviq: argh yes, I looked at the vivid build, sorry [11:11] pitti, so yeah, if it's not in yakkety, can you please remove the PPA build for s390x and I'll work to add a B-D to ubuntu-system-settings to prevent it from building on s390x [11:14] Saviq: done [11:14] pitti, thanks [11:16] I retried unity8 against all of -proposed in ubuntu several times, doesn't help [11:16] the qmlui tests keep being broken [11:26] pitti: that is apparently https://launchpad.net/bugs/1607686 [11:26] Launchpad bug 1607686 in unity8 (Ubuntu) "testWizard and testShellWithPin crashes on yakkety-proposed with Qt 5.6.1" [High,In progress] [11:27] Laney: ah, thanks [11:29] pitti: yep unity8 should be fixed soon, then I'd just need a word from Kubuntu people that they are ok overriding their remaining failures (functionally no-one has complained about anything, even though people are running it) [11:30] it'd be nice to get forward with the proposed migration [11:41] doko, sure I'll take a look [11:51] doko, I'm not seeing a python-taskflow ftbfs. did you mean python-eventlet? [11:51] coreycb, http://people.ubuntuwire.org/~wgrant/rebuild-ftbfs-test/test-rebuild-20160705-gcc6-yakkety.html [11:52] doko, gotcha, thanks [11:52] coreycb, but let me retry ... [11:54] doko, I just hit the same errors on amd64. I'll work on fixing that up. [12:05] Is this the spot to ask packaing questions? I'm trying to package a cmake based library. However dpkg-gensymbols complains about new symbols appearing, such as "_ZNSt7__cxx1115basic_stringbufIcSt11char_traitsIcESaIcEED0Ev@Base 2.1.2". I suppose there is a dep missing, and I tried googling it, but without any results. Is this known to anyone here? [12:09] sveinse: see dpkg-gensymbols. [12:26] rbasak: Yes. I have no clue thou if I can simply add it to the debian/*.symbols file without any side effects. Should I? [12:30] pitti: update_excuses stopped updating again [12:57] sorry if before I wrote in the middle of another discussion. Now I retry to expose my problem [12:57] I am running a Xubuntu 14.04 system. I found some mismatches between the number of "users" printed in the output of "uptime" and the *actual* number of users locally connected to the system [12:58] here I pasted the outputs I got: http://paste.ubuntu.com/21877593/ [12:58] as you can see, both w and uptime indicate that 5 users are connected, but both w and who show just 2 users [12:59] and there are actually 2 users: one who made a login from the GUI, and one from ssh [12:59] I didn't use nor screen neither tmux [13:00] thanks to the help of users in the #ubuntu channel, Ubuntu 16.04 seems to behave in a different way. So I would like to ask: why is there this mismatch? [13:01] (I mean that probably in Ubuntu 16.04 both w and who would have correctly shown 5 user logins) [13:04] there can be some errors in the implementation of uptime in *ubuntu 14.04? [13:12] rockyh: This is still really a question for #ubuntu; could you re-join there and we can discuss it? === _salem is now known as salem_ === mhall119_ is now known as mhall119 [13:32] Odd_Bloke: they suggested me to write here [13:33] Odd_Bloke: anyway, sure, I just re-joined #ubuntu [13:46] looks like britney is crashing [13:50] doko: hello, thought i would ask here instead of empty #ubuntu-toolchain -> llvm-3.8 has recently landed in trusty-updates, but it seems that openmp is not in? [13:51] doko: when i try to compile anything with omp.h & -fopenmp it fails, and there is no libomp-dev like in xenial present [13:52] zaytsev, please ask tjaalton about llvm in trusty [13:53] doko: thank you! tjaalton ^^^ any comments? thanks! [13:56] tjaalton: i see in rules you left the following in - --with-clang-default-openmp-runtime=libomp - but libomp hasn't been backported / nowhere to be found :-/ [14:01] zaytsev: ok, need to drop that then [14:01] pitti: did you change anything in britney today? [14:03] tjaalton: means we ain't gonna get omp? fair enough... i'll turn off omp for llvm on our trusty boxes then. [14:04] zaytsev: it's backported only for hwe lts stack (mesa) [14:09] tjaalton: sure, i understand. still.. do you think maybe hadcoding libgomp as a default runtime could work? [14:10] no idea [14:11] oh britney, why you so sad? [14:17] Laney, I may have a hack for unity-greeter's prompt box. Doesn't seem to work for the session chooser, but looking at that next [14:17] mterry: !!!! [14:18] hacks are good at this stage [14:19] Laney, it basically sets the prompt box's "resize_mode", which is a deprecated value. And if that deprecated value is set, gtk skips the new resize logic patch you linked [14:19] "might introduce obscure bugs if used" [14:19] haha [14:20] so you have to queue a resize yourself manually or something? [14:23] Laney, no... I'm settingn the resize_mode to "immediate" and it seems to work without other changes [14:23] Laney, one line hack so far -- but session chooser may be another thing [14:42] Laney, jbicha: eek, thanks for pointing out; fixed [14:43] Laney: yes, I did an upstream update to fix the missing architecture build excuses in the HTML [14:46] pitti: I already fixed it [14:46] hopefully [14:46] it's copying stuff now [14:46] oh, you force pushed over that [14:46] Laney: sorry, mid-air collision [14:47] it wasn't mid air - that was pushed and already running [14:47] I fixed the iteration over block-bugs [14:47] same [14:47] oh well [14:47] it's the one thing we don't test, as we don't have a mock for Launchpad -- I guess it's time to write one [14:47] suggest pull before push next time [14:48] http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html just updated [14:48] Laney: yeah, sorry, that was too quick [14:49] pitti: you can just supply a Blocks file for testing? [14:49] pitti: also, can you make the runner script run git reset -q please? [14:49] ubuntu-archive's mailbox is getting spammed :) [14:49] sure [14:49] would do it, but EPERM [14:53] Laney: done [14:54] danke! [14:56] mterry: I filed bug #1608908 earler, btw [14:56] bug 1608908 in unity-greeter (Ubuntu) "Some UI elements are invisible with GTK 3.20" [Undecided,New] https://launchpad.net/bugs/1608908 [14:56] nothing new in there, but a place to hang your fixes [14:56] :) [14:56] k [15:06] doko, hi, please take a look at the patch changes https://paste.debian.net/plain/786589 [15:08] Laney: I did a round of cleaning up on all the armhf and s390x boxes, and refreshed the armhf lxd ones as well; should hopefully hold up until end of next week :) [15:09] Laney: i. e. on armhf lxc sometimes leaks containers, or lxc-start/lxc-stop gets stuck etc., but it's ok to pile up a few of those [15:14] infinity: hello & good morning! any movement on bug 1605795 since we last spoke? just checking in :) [15:14] bug 1605795 in livecd-rootfs (Ubuntu) "[SRU] livecd-rootfs ubuntu-cpc vagrant image builder" [Undecided,Confirmed] https://launchpad.net/bugs/1605795 [15:14] Odd_Bloke: ^^ [15:17] pitti: ah, nice - how do you see the lxd stuff? [15:17] I'm on the controller [15:17] * Laney hasn't used that before [15:29] Laney: it's autopkgtest-lxd-worker/0 [15:29] Laney: instead of cloud-worker [15:30] pitti: you just kill the workers in the same way? [15:31] Laney: yes, other than that they use a remote lxd instead of ssh they are the same [15:31] ok [15:31] I thought you were saying you went a poked lxd directly [15:31] check "lxc remote list" and e. g. "lxc list lxd-armhf-236: [15:33] Laney: the remotes are still horribly brittle though, so I don't actually expect this to get much work done [15:33] Laney: so this is still an experiment [15:33] but you still have the static lxc ones, so that should be okay [15:35] right [15:42] * pitti waves good bye, holiday o'clock -- see you all in two weeks! [15:43] see you pitti, have fun! [15:43] thanks! [15:43] * pitti disconnects from the hive, argh === salem_ is now known as _salem [16:00] infinity, kees, stgraber: TB meeting? [16:00] slangasek: thanks for the reminder === _salem is now known as salem_ [16:42] slangasek: Ungh. A little timezone challenged this morning. [17:18] Laney, so... I have a fix for the prompt & the session list. But after going into the session list and back out again, there is a tiny visual glitch when switching between users. I'm inclined to not care about that. Haven't had luck fixing it yet. Might just push as-is, especially if gtk3.20 is about to land === msbrown-afk is now known as msbrown [20:18] mterry: That's fine, it sounds small enough to deal with later if necessary [20:18] thanks a bunch for figuring it out [20:19] Laney, uploaded in -proposed already [20:19] you're welcome! didn't want people to yell at me for a broken unity-greeter :P [20:20] GTK's blocked on an MIR and one or two other uploads anyway, but yeah :) [20:21] working on that MIR right now [20:22] * Laney remembers to subscribe the team [20:22] I unsubscribed it when I un-MIRed [20:24] Laney: did you see that vte is stuck in proposed because pcre2 is in universe? [20:24] Laney: but it looks like the dependency isn't even used now https://git.gnome.org/browse/vte/commit/?h=vte-0-44&id=ce94be [20:25] jbicha: nope [20:30] yeah, it builds fine without pcre2 [20:30] I know, it's just new API [20:31] need to turn it off in gnome-terminal too [20:36] Laney: could you do a no-change rebuild of webkit2gtk for gtk 3.20? [20:37] it's on the list [20:38] guess I could do it now though [20:42] jbicha: I did it [20:42] now before you give me any additional work, goodnight! [20:42] :) [20:43] Laney: see ya :) === salem_ is now known as _salem