[00:04] <jbicha> TheMuso: thanks
=== JanC is now known as Guest98943
=== JanC_ is now known as JanC
[02:16] <TheMuso> jbicha: Ok works fine here, will upload.
[02:17] <jbicha> TheMuso: great! I appreciate it
[02:18] <TheMuso> jbicha: You're welcome.
[04:48] <pitti> Good morning
[04:48] <pitti> Laney: many thanks for quick-fixing the graphical-pre thing!
[04:49] <pitti> that didn't break the normal ubuntu desktop hard enough to notice :(
[04:52] <hikiko> hi
[05:41] <pitti> tedg: hey Ted, how are you? What's the current status of https://requests.ci-train.ubuntu.com/#/ticket/1710 ? it seems unity got dropped there?
[05:42] <pitti> tedg: if we replace "manual" overrides with "exec sleep infinity" to keep stub upstart jobs, we might land this in smaller chunks
[08:01] <Laney> morning!
[08:02] <pitti> hey Laney, good morning
[08:02] <Laney> hey pitti
[08:02] <seb128> good morning desktopers
[08:02] <Laney> you're welcome on the systemd
[08:02] <pitti> bonjour seb128 !
[08:02] <Laney> how's it oging?
[08:02] <Laney> hi seb128!
[08:02] <pitti> Laney: it's oging okay, thanks; my cold is getting better, had a calm weekend
[08:02] <pitti> how about you?
[08:04] <Laney> seb128: wie gehts?
[08:05] <Laney> pitti: I'm good - climbing, visited a new hipster coffee shop and went to see a singer at the pub
[08:05] <Laney> oh, also replaced my shoes that had a hole in them finally /o\
[08:05]  * Laney put the old ones in a bin outside the shop
[08:07] <seb128> Laney, gut! you? had a good w.e?
[08:08] <seb128> had a nice saturday, got some nice food&beer at the Rotterdam market and did a bbq in the evening
[08:09] <seb128> then a relaxing sunday afternoon in France (after driving in the morning)
[08:09] <Laney> nice
[08:09] <Laney> how long are you in france for?
[08:10] <seb128> this week and a few days next probably, didn't decide yet when I drive back but at least staying for the week and w.e
[08:11] <Laney> sweet
[08:11] <seb128> yeah, and they forecast real summer which is nice!
[08:11] <pitti> here too, from Wed on (but even today it's sunny again, just not really warm yet)
[08:11] <seb128> like 28-30°C sunny and not too windy
[08:11] <Laney> that's going to be the last iteration
[08:11] <Laney> take some days off ;-)
[08:12] <seb128> here it's 25°C today and should be 29°C and very nice starting tomorrow
[08:12] <seb128> yeah, I've a swap day I kept for this week (worked aug 15)
[08:12] <seb128> I'm probably going to do an extended w.e ;-)
[08:13] <Laney> nice
[08:13] <Laney> uk has bank holiday next monday
[08:13] <Laney> join us
[08:13] <seb128> ah, nice
[08:13] <seb128> what's the occasion?
[08:14] <Laney> don't actually think there is one
[08:14] <Laney> it's called "late summer bank holiday"
[08:15] <Laney> we're patriotic over here
[08:16] <Laney> last one until christmas day though
[08:16] <seb128> yeah, we don't have many left either
[08:17] <davmor2> Laney: No we just worship banks obviously
[08:17] <pitti> we have one remaining one on Nov 1
[08:17] <Laney> davmor2: yeah - when they holiday, we holiday - because there's no way you can do any work when the banks are shut
[08:18] <seb128> we have nov 1 and 11 (armistice from ww1)
[08:18] <davmor2> Laney: if we were Patriotic we would have Saint Georges Day off :P
[08:19] <davmor2> seb128: we have 2 minutes Silence at 11 on the 11/11
[08:20] <seb128> you should have 11
[08:20] <seb128> :-)
[08:20] <davmor2> seb128: hahaha
[08:25] <Laney> brb
[08:29] <Sweet5hark1> moin
[08:32] <seb128> hey Sweet5hark1, how are you? had a good w.e?
[08:32] <andyrock> morning all
[08:33] <Sweet5hark1> davmor2: well, dont try that in south-western germany. 11/11 on 11:11 is the start of the "foolish season" there: the start of preparations for carneval!
[08:33] <willcooke> o/ (sort of)
[08:36] <Sweet5hark1> seb128: weekend was quite well: some BBQ at my bothers place saturday, some do-nothing-day on sunday. Friday evening I watched an italian entry of the  "Schlechteste Filme aller Zeiten" (worst movies of all times) on a local TV station, which was excellent: It was not only boring bad, it was the kind of hilarious and campy bad that is enjoyable to watch, even with just mild levels of intoxination ...
[08:36] <seb128> heh
[08:37] <seb128> hey willcooke!
[08:37] <seb128> hey andyrock
[08:37] <seb128> had a good w.e?
[08:37]  * Sweet5hark1 prepares a hot toddy for willcooke.
[08:37] <willcooke> :) thx
[08:37]  * andyrock went to a wedding
[08:37] <willcooke> Figured I would check on email and that, and then probably crash out for a couple more hours
[08:37]  * andyrock is still full
[08:38] <andyrock> seb128: you never had this problem? https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1599516
[08:38] <ubot5> Launchpad bug 1599516 in unity (Ubuntu) "No accented letters in window title bars, dash and hud." [Medium,New]
[08:41] <seb128> I feel like I missed some info
[08:42] <seb128> willcooke, sick? or just tired?
[08:42] <seb128> andyrock, no I didn't, it's weird, I wonder if they use a non utf8 locale or something
[08:42] <willcooke> seb128, sick
[08:42] <seb128> oh, get better!
[08:43] <willcooke> seb128, thanks!  Only a cold, but the usual temperature, sneezing, coughing, etc
[08:43] <seb128> :-(
[08:43] <andyrock> seb128: https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1599516/comments/11
[08:43] <ubot5> Launchpad bug 1599516 in unity (Ubuntu) "No accented letters in window title bars, dash and hud." [Medium,New]
[08:44] <andyrock> from this comments it seems that they have the correct locale
[08:44] <willcooke> actually, I need to go and buy some proper tissues instead of using toilet paper or kitchen paper
[08:45] <seb128> andyrock, I would ask for the content of /proc/`pidof compiz`/environ and if that happens in a guest session
[08:46] <seb128> andyrock, try asking for ~/.cache/upstart/unity7.log also
[08:50] <Sweet5hark1> seb128: any idea what I can do about the LibreOffice ftbfs? Because I am out of ideas and I think Im blocked.
[08:50] <seb128> Sweet5hark1, disabling that group of test isn't enough?
[08:50] <andyrock> oki
[08:51] <Sweet5hark1> seb128: well, I disabled the hanging test. then it hung on the next one. I might still try to go full geronimo and disable _all_ tests ....
[08:55] <Sweet5hark1> seb128: note that the four test we found hanging from the pstree are completely different kinds of beasts: LibreOfficeKIt, Draw file import, Lotus text file filters, Calc tiled rendering. If they all are legitimately failing, that means essentially that nothing works.
[08:56] <Laney> sbuild doesn't reproduce it?
[09:00] <Sweet5hark1> Laney: ah, ok that is another thing I can try indeed: I did a pbuild on yakkety w/o proposed and a "dpkg-buildpackage" from a terminal in a VM with yakkety-proposed, but I didnt yet try a sbuild/pbuild with yakkety-proposed yet ...
[09:00] <seb128> try to sbuild
[09:00] <Laney> that's the closest you can easily get to the launchpad environment
[09:00] <seb128> I can retry the current version, sometime it works
[09:00] <Laney> also...
[09:00] <seb128> we might be able to get that one in on retries
[09:00] <Laney> ...
[09:01] <Laney> can we give up on firefox?
[09:01] <Laney> it's beta 1 for some flavours this week
[09:01] <seb128> infinity said he would look at it?
[09:01] <Laney> people keep saying that
[09:01] <Laney> can we just remove it and then when someone fixes it it'll come back?
[09:01] <Sweet5hark1> Laney: do you know by chance what the host OS for the sbuild is on launchpad?
[09:02] <Laney> nein, but you might get some clues from the top of a build log
[09:02] <Laney> chance it varies across architecture too
[09:02] <seb128> Laney, it's not as easy as deleting it apparently, need to look at rdepends and seeds from what other said
[09:02] <seb128> we could delete and see what issues it creates if any
[09:05]  * Sweet5hark1 is slightly nervous about the fact that german government recommends having at least 30 days of food reserves in your home again now. First time since the end of the cold war.
[09:07] <Laney> i'm checking the rdeps quicklyi
[09:08] <seb128> Sweet5hark1, why do they do that?
[09:08] <seb128> Laney, thanks
[09:09] <seb128> hum
[09:09] <seb128> is yakkety-proposed b0rked?
[09:09] <seb128> https://launchpadlibrarian.net/280303103/buildlog_ubuntu-yakkety-amd64.libreoffice_1%3A5.2.0-0ubuntu1_BUILDING.txt.gz failed with
[09:09] <seb128> "/bin/bash: python: command not found"
[09:12] <Laney> interesting
[09:14] <infinity> Laney: I said it and I meant it, last week was just hellish with the Qt/binutils/kernel entanglement.
[09:15] <Laney> infinity: I know you mean it - but I also don't want to have Beta 1 with the knackered version we have now
[09:16] <Laney> Fixing it on those architectures doesn't need to be a stressful thing done under time pressure
[09:16] <infinity> Does it matter?  It's an opt-in milestone.
[09:16] <Sweet5hark1> seb128: well, they say for emergencies for the civil population like "natural desasters", animal epidemics, industrial accidents (e.g. nuclear power plants), terrorism. war is not explicitly mentioned, only in euphemisms ("territorial attack", "conventonal defense" ...)
[09:16] <infinity> But meh, force it if you want.  Don't remove the other binaries, or the world explodes.
[09:16] <Laney> It's also quite broken in the yakkety-release for everyone. :P
[09:17] <Laney> Ta
[09:18] <Sweet5hark1> Laney, infinity: python.qt/binutils/kernel fun? so yakkety is really the yolo release?
[09:18] <Laney> seb128: seems there's no build-depends on python
[09:19] <happyaron> seb128: mind to have a look at nm-applet again? yakkety branch (master) is updated
[09:19] <seb128> Laney, was it part of the base until now? I just wonder what changed between a week ago and today
[09:19] <seb128> happyaron, hey, sure, thanks for updating it
[09:20] <seb128> shrug, launchpad timeouts are getting annoying nowadays
[09:23] <Trevinho> good late morning guys
[09:24] <Laney> seb128: hmm, something definitely did change
[09:24] <seb128> hey Trevinho! how are you? had a good w.e?
[09:24] <seb128> Laney, yeah, I'm just wondering what...
[09:24] <seb128> is python3 supposed to provide a "python" binary?
[09:25] <Trevinho> hey seb128, I'm ok thanks... Nice weekend, some rest on saturday and swimming pool day yesterday... Ended with a never-ending dinner :-)
[09:25] <Trevinho> (this is a loop...)
[09:25] <seb128> nice
[09:25] <seb128> still at dinner then? ;-)
[09:25] <seb128> was it a date at least? :p
[09:25] <Trevinho> ahah, lol... sort of
[09:26] <Trevinho> na, na... It was a date with so much local food :-D
[09:26] <Trevinho> Generally better than girls... It gives you pleasure, without the hassle of speaking :-D
[09:27] <Laney> certainly not
[09:27]  * Trevinho jokes
[09:27] <Laney> it looks like libglib2.0-dev used to cause python 2 to be installed
[09:27] <Laney> but now it switched to python3 that's not the case
[09:28] <seb128> I see
[09:28] <seb128> so libreoffice bug
[09:29] <seb128> Sweet5hark1, ^ you need to build-depends on "python" yourself if you used it
[09:29] <seb128> use
[09:29] <Laney> yeah python isn't part of the minimal set
[09:29] <Laney> perl, however... :)
[09:29] <seb128> haha
[09:29] <seb128> Trevinho, Italian food p0rn :-)
[09:30] <Trevinho> ahah
[09:30]  * Laney takes the advice and force-hints firefox
[09:31] <seb128> thanks Laney
[09:31] <Laney> can haz fixed themes
[09:33] <ochosi> Laney: themes - fixes - fill me in plzzz? :)
[09:33] <Laney> firefox
[09:33] <Laney> it had a broken version in the release
[09:34] <Sweet5hark1> seb128: hmm? /me does not quite get the context wrt b-d on python and libreoffice bug. is this supposed to be related to the hang or because of unrelated yakkety-proposed changes in the mean time?
[09:34] <ochosi> Laney: oh ok, i'm still on xenial with gtk3.20 and with one of the more recent FF updates it started to look ok again (actual scrollbars and stuff)
[09:34] <seb128> Sweet5hark1, see what L_aney just wrote and the build log I gave a bit earlier
[09:34] <Laney> yeh
[09:35] <seb128> Sweet5hark1, but basically libreoffice build retries failed on missing python, glib used to pull it in for you and stopped doing so
[09:35] <seb128> unrelated to the tests hang
[09:37] <Sweet5hark1> seb128: ah, seeing it now, thx. thats odd: why does that build recreate ./deabian/control at all? curious!
[09:38] <seb128> Sweet5hark1, I guess it's your debian/rules doing it?
[09:38] <seb128> the log also has that, unsure if that's an issue worth looking at
[09:38] <seb128> "dh_listpackages: -s/--same-arch is deprecated; please use -a/--arch instead
[09:38] <seb128> basename: missing operand
[09:38] <seb128> Try 'basename --help' for more information."
[09:39] <seb128> debian/rules:763: SYSTEM_GCJ_VERSION empty, falling back to SYSTEM_GCC_VERSION
[09:39] <seb128> there is several of those basename missing operand errors
[09:40] <Sweet5hark1> seb128: yes, it is ./debian/rules doing it, but it shouldnt obviously ...
[09:42] <seb128> Laney, the unity-greeter s390x binary in yakkety-proposed needs to be deleted right?
[09:43] <seb128> "missing build on s390x: unity-greeter (from 16.10.2.1) "
[09:43] <Laney> yes if it wasn't already
[09:43] <Laney> .2 should have no build there since I added a build-dep on upstart
[09:44] <seb128> rmadison seems to suggest it was not
[09:44] <Laney> ok, please do, thanks!
[09:45] <seb128> done, yw!
[09:45] <seb128> hum
[09:45] <seb128> pcre3 migrated
[09:45] <seb128> but it makes the autopkgtests for glib fail, only works with the proposed version of glib that didn't migrate
[09:46] <seb128> is that supposed to happen like that?
[09:46] <seb128> I never remember what is tested against what...
[09:46] <Laney> shouldn't do
[09:47] <Laney> unless it was run against all of proposed I guess
[09:47] <seb128> is that an hammer people start abusing?
[09:49] <Laney> not in this case
[09:49] <Laney> http://autopkgtest.ubuntu.com/packages/g/glib2.0/yakkety/amd64/
[09:50] <seb128> I guess it's because pcre3 doesn't have tests itself?
[09:50] <seb128> it was just reverse hitting the glib ones
[09:51] <seb128> but nothing forced them to migrate together
[09:51] <Laney> I think it doesn't track the versions of reverse dependencies that worked and didn't work
[09:53] <seb128> hum, k
[09:54] <Laney> could be wrong though
[09:55] <Laney> Something like ideally it would have seen that the new pcre3 only worked with glib2.0 2.49.5-1 and then added a fake dependency into britney
[09:55] <Laney> That probably has lots of hairy cases I'm not thinking of
[09:59] <seb128> yeah
=== hikiko is now known as hikiko|ln
[10:01] <robert_ancell> Laney, doesn't Architecture: [!s390x] solve the u-g build issue?
[10:02] <Laney> hey robert_ancell
[10:02] <robert_ancell> Laney, hi!
[10:03] <seb128> hey robert_ancell
[10:03] <Laney> if you add the build-dep then it automatically builds again when someone fixes the package
[10:03] <Laney> hardcoded lists like that are more difficult to undo
[10:03] <robert_ancell> ah
=== vrruiz_ is now known as rvr
[10:25] <pitti> Laney, seb128: gnome-terminal used to recognize "LP: #XXXXX" patterns and made them clickable, it seems to now have lost that; is that GTK or gnome-terminal or vte? known issue?
[10:27] <Laney> it's a patch in gnome-terminal
[10:31] <pitti> so I guess a regression in https://launchpad.net/ubuntu/+source/gnome-terminal/3.20.2-1ubuntu1, although the merge changelog still says that 60_add_lp_handler.patch exists
[10:31] <pitti> darkxst: ^
[10:33] <Laney> nod, sorry for the problem
[10:34] <pitti> no worries, not a biggie; but I only just now recognize how often I use this :)
=== hikiko|ln is now known as hikiko
[11:04] <happyaron> seb128: mind to retry this build? https://launchpad.net/ubuntu/+source/network-manager/1.2.2-0ubuntu0.16.04.1/+build/10649216
[11:05] <seb128> happyaron, done
[11:05] <happyaron> thx
[11:06] <seb128> yw!
[11:31] <Laney> pitti: think I've fixed it, all for you
[11:31] <Laney> ;)
[11:35] <Laney> lp: #1234 lP: #1234 LP: #1234
[11:35] <ubot5> Launchpad bug 1234 in Launchpad itself "Gina is an unmaintainable mess of command line options, environment variables and shell scripts" [Medium,Fix released] https://launchpad.net/bugs/1234
[11:35] <Laney> haha
[11:38]  * xnox shall use that next "unmaintainable mess"
[11:41] <mdeslaur> sounds like most tools we use
[11:50] <Sweet5hark1> *cough*
[11:57]  * Laney forces harder
[12:09]  * Sweet5hark1 is an old neckbeard that uses this newfangled sbuild for the first time instead of his heavily scripted pbuilder ...
[12:09] <Sweet5hark1> .oO(Kids these days)
[12:13] <davmor2> Sweet5hark1: they need to get off my lawn /me waggles stick in general direction of kids
[12:24] <pitti> Laney: *hug*, merci beaucoup !
[12:49] <xnox> Sweet5hark1, hehe. It's been years since i've purged all of my configs and rely solely on $ mk-sbuild foo; sbuild -d foo
[12:50] <xnox> Sweet5hark1, however, it still needs a few tiny tweaks in .sbuildrc for edge cases: $resolve_alternatives = 1;
[12:50] <xnox> and
[12:50] <xnox> $build_environment = { 'NO_PKG_MANGLE' => '1', 'DEB_BUILD_OPTIONS' => 'parallel=12', 'HOME' => '/build/' };
[12:52] <Sweet5hark1> xnox: yeah, to really use sbuild Id need to glue it into my existing wrappers. so far that was "never touch a running system" though ...
[12:56] <donald>  bonjour je veux installer un multiboot.  j'ai déjà installé un ubuntu. Je veux installer un kali à la suite. Je suis en train de partitionner manuellement avec le guide https://openclassrooms.com/courses/reprenez-le-controle-a-l-aide-de-linux/partitionner-son-disque. Quel point de montage est-ce que je dois choisir pour la partition qui contiendra  le systeme de kali. est-ce bien "/" ?
[13:00] <pitti> donald: désolé, on parle anglais ici, et cette chaîne est pour ubuntu, pas pour kali; mais '/' semble correct pour la partition de système
[13:10] <seb128> Laney, seems your firefox hammer worked ;-)
[13:22] <pitti> oh, nice! was that a "slash away powerpc binaries" hammer? :-)
[13:25] <Laney> boooom
[13:25] <Laney> pitti: I deployed a force hint
[13:25] <pitti> "use the force, Luke^WLaney!"
[13:29] <Trevinho> :D
=== czajkows1i is now known as czajkowski
[14:02] <MikeB__> I was referred to this channel for possible help with a desktop problem that showed up on my Oracle VirtualBox VM after upgrading my Xenial last week.  I don't really know exactly what package or version caused the problem.
[14:03] <MikeB__> After the upgrade, all new windows showed up partially hidden behind the launcher bar on the left and the status line on the top.
[14:04] <MikeB__> Before the upgrade, new windows always appeared fully - i.e. not hidden behind launcher bar and top status line.
[14:23] <mhall119> willcooke: ping, any word on a testable Unity 8 ISO?
[14:34] <seb128> mhall119, that's not something being worked
[14:35] <seb128> MikeB__, can you pinpoint what update caused the issue? like start from a working image and/or downgrade?
[14:37] <seb128> pitti is doing cowboy style landing to u-s-d without using mps/reviews/ci ;-)
[14:38] <pitti> Yeeehaw!
[14:38] <seb128> :-)
[14:38] <MikeB__> @seb128 - I'll try. Can you give me the appropriate downgrade command - I've never downgraded before.
[14:38] <meetingology> MikeB__: Error: "seb128" is not a valid command.
[14:39] <seb128> pitti, just for the record desktop team is reviewer for that one so usually it doesn't get stucked in no-landing-land that sometime happen on e.g indicator
[14:39] <seb128> pitti, it's always good to have peer reviews, but I had a look at the launchpad diff and it looked fine so no worry for that one
[14:39] <pitti> seb128: ok, noted for next time; I noticed that it doesn't use double/triple landing, so just pushing to bzr and uploading to yakkety seems to not disturb the train
[14:40] <seb128> right, it just hijacked the review process
[14:40] <seb128> which is fine for easy changes but there is still some value to have somebody else to look and ack
[14:42] <seb128> MikeB__, get the debs from e.g https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-063/+build/10465506 and dpkg -i those
[14:56] <seb128> jbicha, Laney, is any of you looking at the cjs/isenkram test regressions that are blocking gtk in yakkety-proposed?
[14:59] <jbicha> I filed https://bugs.debian.org/834374
[14:59] <ubot5> Debian bug 834374 in cjs "cjs: Failing autopkgtests" [Important,Open]
[15:00] <seb128> thanks
[15:00] <seb128> did you look at isenkram as well?
[15:06] <jbicha> no, maybe that's pygobject? but isenkram hasn't been autopkgtested on ci.debian.net in a few weeks
[15:08] <jbicha> I don't know enough about what's going on there to do much
[15:14] <ricotz> jbicha, hi, I guess mutter should be blacklisted to avoid it gets auto-synced?
[15:14] <jbicha> ricotz: we are past Debian Import Freeze and autosync has been disabled for yakkety so I think we're good
[15:15] <ricotz> jbicha, ah, good
[15:23] <superm1> ximion: Laney okay well whichever of you guys want to do the backport, the patches series should be in good shape for that issue now
[15:47] <seb128> pitti, did you see bug #1615230?
[15:47] <ubot5> bug 1615230 in systemd (Ubuntu) "Policykit broken with latest systemd" [Undecided,Confirmed] https://launchpad.net/bugs/1615230
[16:00] <jbicha> Laney: do we want to remove --enable-debug from gnome-terminal now (it adds the Help>Inspector item) or is still theme work being done that would be helped by that feature?
[16:09] <Laney> jbicha: Not right now, but is it a problem to leave it?
[16:11] <jbicha> Laney: I think we want to remove it before release (so maybe by UIF) because it's not something ordinary terminal users need
[16:13] <Laney> If you care that much then do it - I'm not bothered about it being there personally
[16:14] <Laney> If you want to do something nice (upstream), then maybe make it look at the gsettings key that gtk provides instead of being a compile time flag
[16:17] <seb128> andyrock, Trevinho, is that known that yakkety daily in a vm displays the keybinding summary but that you can't close it with the mouse (clicking the X doesn't work, clicks go through to desktop/nautilus)
[16:17] <seb128> also nautilus icons are overlapping
[16:17] <seb128> but that's likely a new nautilus issue
[16:18] <seb128> Laney, unity-greeter migrated \o/ small side effect, it makes kylin-greeter be on component mismatch, I guess that's due to s390x and trying to find a greeter alternative
[16:19] <jbicha> Laney: I think the problem is gnome-terminal's unique server/client thing
[16:20] <Laney> seb128: yeah nice
[16:20] <Laney> how do you want to fix that kylin thing?
[16:20] <Laney> jbicha: what problem?
[16:21] <jbicha> here, I'll file a bug and see what chpe says about g-terminal
[16:22] <Laney> yeah I would suggest asking before starting any work
[16:22] <seb128> Laney, unsure, do we have another greeter in main we could add as alternative? otherwise I guess adding a [!s390x] to the recommends...
[16:22] <jbicha> https://wiki.gnome.org/Apps/Terminal/Debugging
[16:23] <Laney> grim
[16:24] <jbicha> yeah, he has to create wiki pages for all the stuff people repeatedly file bugs about
[16:24] <Laney> jbicha: Don't see how any of that gets in the way of using this key to show or hide the menu item
[16:25] <jbicha> ok
[16:25] <Laney> I think anyway
[16:25] <Laney> just an idea to avoid having to turn it off
[16:26] <Laney> :)
[16:27] <jbicha> I did previously file https://bugzilla.gnome.org/767665
[16:27] <ubot5> Gnome bug 767665 in general "can't run GtkInspector with gnome-terminal" [Normal,Resolved: notabug]
[16:29] <Laney> I suppose an argument against is that then you have to enable-debug and set the key
[16:29]  * Laney shrugs
[17:27] <pitti> seb128: I did see it, just didn't look into it yet (WFM)
[18:22] <Sweet5hark1> hey guys, I have some good news and some bad news!
[18:22] <Sweet5hark1> The good news: I can reproduce the hang locally.
[18:23] <davmor2> Sweet5hark1: bad news you can't access the system anymore to find out what hung?
[18:23] <Sweet5hark1> The bad news: it doesnt make much sense at all.
[18:24] <Sweet5hark1> at least the back traces dont.
[18:25] <Sweet5hark1> ah, hmmm, they start to make some sense ...
[18:25] <Sweet5hark1> ... or at least some trail to follow ....
[18:28] <Sweet5hark1> which is: cups is evil!
[18:28] <Sweet5hark1> (maybe)
[18:57] <seb128> Sweet5hark1, nice to see you can reproduce/have a system to debug!
[18:59] <Sweet5hark1> seb128: Im seeing multiple cups subprocesses busy waitiing on pipes. the backtraces are still nonsense, so I wonder how they got there ...
[18:59] <seb128> Sweet5hark1, if it's easy to reproduce maybe downgrade cups and see if it goes away
[19:00] <seb128> you might also want to ask tkamppeter if there is a known cups issue
[19:00] <Sweet5hark1> seb128: hah! there is a race condition somewhere in cups!
[19:00] <seb128> oh?
[19:01] <seb128> so libreoffice is not only exercising toolchains? ;-)
[19:02] <tkamppeter> To me there is no issue known of CUPS hanging, perhaps some problem of 2.2.x?
[19:02] <tkamppeter> seb128, Sweet5hark1 ^^
[19:02] <Sweet5hark1> there are multiple cppunittesters busy waiting on pipes or something. once I kill the (sub)process, the build continues ...
[19:02] <seb128> tkamppeter, yeah, that started recently so if it's a cups issue it's likely the 2.2 update
[19:03] <seb128> Sweet5hark1, if you can trigger it easy try to downgrade cups
[19:03] <tkamppeter> seb128, Sweet5hark1: Or do you have the hang on shutdown, with Xenial, then it can be bug 1579905.
[19:03] <ubot5> bug 1579905 in cups-filters (Ubuntu Xenial) "cups-browsed causes shutdown hang/delay in Ubuntu 16.04" [Medium,Fix committed] https://launchpad.net/bugs/1579905
[19:03] <Sweet5hark1> seb128: yep
[19:04] <Sweet5hark1> tkamppeter: well, I have the hang in yakkety sbuild running on xenial ...
[19:04] <tkamppeter> Sweet5hark1, what does mean yakkety sbuild?
[19:06] <tkamppeter> Sweet5hark1, seb128, in case of a CUPS 2.2.x issue we need to do an upstream bug report. So as much info as possible is needed from you, see https://wiki.ubuntu.com/DebuggingPrintingProblems
[19:06] <Sweet5hark1> tkamppeter: well, Im using sbuild to build libreoffice for yakkety on a xenial host. When I build libreoffice on a yakkety VM it does not hang. only in the sbuild (and on launchpad)
[19:07] <tkamppeter> Sweet5hark1, is the hang you are observing happening in the middle of the session without special reason? Or when trying to print? Or on shutdown of the system?
[19:08] <tkamppeter> Does the sbuild involve starting or shutting down a VM or something similar? Or does it try to shut down CUPS and or cups-browsed?
[19:08] <Sweet5hark1> tkamppeter: its happening in various unittests, none of them should try to print. but they might do things like enumerate printers and such ....
[19:09] <Sweet5hark1> sbuild is essentially a chroot.
[19:09] <Sweet5hark1> tkamppeter, seb128: http://pastebin.com/YxRPmHEW <- this is the bt of the hanging process ...
[19:10] <seb128> dbg would be nice there I guess
[19:12] <Sweet5hark1> and the hanging process has multiple (~2-4) child processes, here is a bt of them: http://pastebin.com/hfTn6Vv8
[19:13] <tkamppeter> Sweet5hark1, perhaps some debug packages (of both cups and LO) should be installed into that chroot?
[19:14] <Sweet5hark1> so ... something called "CUPSManager cups" spawns 2-4 processes and both that process and one of the childs are busy looping/waiting on pipes. The child is in "ppdCollect2 () from /usr/lib/x86_64-linux-gnu/libcups.so.2"
[19:15] <tkamppeter> Sweet5hark1, it looks like that some functions of libcups are called to get info about available printers and their PPD files (printer properties).
[19:16] <Sweet5hark1> tkamppeter: the question is: does this only happen when I build from scratch? or also when I just start one unittest in the sbuild ... If its the former, it takes hours to get back there ....
[19:16] <tkamppeter> Sweet5hark1, I do not know any "CUPSManager" utility or so.
[19:17] <tkamppeter> Sweet5hark1, in a regular environment such a thing never happened to me. sbuild I never used. I use VMs to build for different Ubuntu versions.
[19:17] <Sweet5hark1> tkamppeter: apparently the thread is named as such: http://pastebin.com/YxRPmHEW <- line 1
[19:18] <Sweet5hark1> tkamppeter: heh, launchpad uses sbuild essentially ;). So libreoffice unittests need to work in the sbuild ...
[19:19] <tkamppeter> Sweet5hark1, googling for CUPSManager, once tells me that in LibreOffice there is CUPSManager class: https://docs.libreoffice.org/vcl/html/classpsp_1_1CUPSManager.html
[19:19] <tkamppeter> Sweet5hark1, and there exists also a gnome-cups-manager.
[19:21] <tkamppeter> Sweet5hark1, interesting is also that the thread is named "CUPSManager cup" and not "CUPSManager cups".
[19:22] <Sweet5hark1> tkamppeter: This used to work with exact this libreoffice code when we build libreoffice 5.2.0~rc4 (=final) on yakkety on 2016-07-31, so something in libreoffice dependencies changed afterwards and broke it ...
[19:22] <Sweet5hark1> tkamppeter: my bet would be on a cups upload ;)
[19:23] <tkamppeter> Sweet5hark1, perhaps a CUPS error_log (in debug2 mode) could be helpful.
[19:25] <Sweet5hark1> tkamppeter: as said: I dont know yet if it would take me hours to get back here ....
[19:26] <tkamppeter> Sweet5hark1, this is the transition from 2.1.4 to 2.2.0: https://launchpad.net/ubuntu/+source/cups/+changelog
[19:28] <tkamppeter> Sweet5hark1, as I am not able to reproduce the problem, I need to find out how the build/test process communicated with CUPS and how CUPS reacted. A lot of this info lands in /var/log/cups/error_log when CUPS is in debug2 mode ("LogLevel debug2" in /etc/cups/cupsd.conf).
[19:28] <Sweet5hark1> urghs, cups has a lot of binary packages.
[19:32] <tkamppeter> Sweet5hark1, I have looked for already reported CUPS 2.2.x bugs on www.cups.org now and there is nothing related to your problem.
[19:33] <tkamppeter> Sweet5hark1, can it be that in your chroot there is perhaps libcups installed but now cupsd running?
[19:34] <Sweet5hark1> tkamppeter: seems we are in this thread: http://opengrok.libreoffice.org/xref/core/vcl/unx/generic/printer/cupsmgr.cxx#176 <- cut to "CUPSManager cup"
[19:36] <Sweet5hark1> tkamppeter:  and that calls cupsGetDests2() which seems to be what hangs then ... http://opengrok.libreoffice.org/xref/core/vcl/unx/generic/printer/cupsmgr.cxx#222
[19:38] <Sweet5hark1> tkamppeter: there sure isnt a cups running in the chroot. but maybe outside the chroot on the host ...
[19:40] <tkamppeter> Sweet5hark1, and for the cupsGetDests2 this cupsd mudt be accessible. Either via localhost:631 or via a socket: /var/run/cups/cups.sock.
[19:41] <tkamppeter> Sweet5hark1, Note that in an all-default Ubuntu installation CUPS only listens through the socket, meaning that it needs to be available in the chroot.
[19:41] <tkamppeter> Sweet5hark1, you can add "Port 631" to cupsd.conf so that cupsd listens also on localhost:631.
[19:53] <Sweet5hark1> tkamppeter: that doesnt help me on launchpad.
[19:54] <Sweet5hark1> tkamppeter: I think I rather will look into shortcutting the search for printers in unittests so that they dont do any CUPSy things at all.
[19:54] <Sweet5hark1> ... but that is for tomorrow.
[19:54] <Sweet5hark1> <- EOD
[19:55] <tkamppeter> Sweet5hark1, at least it couls help you hen debugging a local sbuild process to find out where it break.
=== mchro- is now known as mchro