larsugood morning!08:21
pittilarsu: hey Lars!08:24
pittihello hikiko, too!08:24
pittipretty lonely here with folks travelling..08:24
larsuindeed. Morgen pitti!08:25
alexarnaudGood morning all08:43
alexarnaudhikiko: Hi08:43
willcookemorning gang08:54
larsuhi alexarnaud and willcooke!08:54
ksamakhoy all09:01
Laneysorry for being late09:10
larsug'day Laney!09:10
Laneygot a friend over to paint in our living room09:10
Laneyhad to show her where all the stuff is09:11
Laneyhey larsu09:11
larsuwhat's she painting?09:12
Laneystarting with the skirting09:13
Laneythen window frames09:13
Laneythen walls09:13
Laneythen if that goes really well, part of the outside wall!09:13
Laneybasically we/I could have done it, but we didn't over 15 months or however long we have been here now09:13
Laneyand she was looking for jobs to do09:13
Laneyseemed fair09:14
larsucool :)09:14
alexarnaudLaney: it's a good deal yeah09:14
pittihey Laney!09:38
Laneyhi there pitti!09:38
Laneywhat's up?09:38
Laneyare you having a good sprint?09:39
pittiLaney: guess what -- handholding LXC workers and making adt-run more robust against more errors :)09:39
pittiLaney: I do, yes! (but in the afternoon/night)09:39
pittithe three guys kept me full-time busy with questions, MPs, setup, etcl09:39
pittibut they all managed to set up the infra in juju-local, and start hacking on it09:40
LaneyI thought that it was going to cover errors, CI train and other related bits too09:41
Laneyseems not?09:41
pittiLaney: not this sprint, this is for p-m09:41
pittiLaney: we'll have 3 more sprints for the other pieces of infra09:41
pittiLaney: today we want to look into a web/SSO based test retry button09:42
LaneyI guess that will be the first dynamic piece on the server09:43
pittiyeah, scary09:44
pittiso far everythign is nice and read-only09:44
pittiLaney: btw, do you know why http://autopkgtest.ubuntu.com/packages/g/glib2.0/xenial/s390x/ is so flappy?09:44
pittiLaney: is that a matter of running against more triggers in -proposed, or just a flaky test and I'll retry until it sticks?09:45
Laneypitti: I didn't really look yet09:45
Laneyis it always the same failure?09:45
pittiLaney: I'm still trying to find the failure; it's not "FAIL:", it's not "ERROR:", goddammit world, agree to some common string to search for09:46
pittiit's not "not ok" either09:46
Laneypitti: "killed by signal"09:47
pittiGLib-GIO:ERROR:/build/glib2.0-dQLvhN/glib2.0-2.47.4/./gio/tests/gdbus-proxy.c:826:fail_test: code should not be reached09:48
pitti# GLib-GIO:ERROR:/build/glib2.0-dQLvhN/glib2.0-2.47.4/./gio/tests/gdbus-proxy.c:826:fail_test: code should not be reached09:48
pitticleaning up pid 685609:48
pittiTest glib/gdbus-proxy.test failed: Child process killed by signal 609:48
pittiI think that's the one09:48
Laneyoh man, gdbus tests09:48
pittifailure in https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-xenial/xenial/s390x/g/glib2.0/20160119_151405@/log.gz looks different09:49
pitti(/usr/lib/glib2.0/installed-tests/glib/gdbus-threading:6884): GLib-GIO-ERROR **: connection had too many refs09:49
pitticleaning up pid 688609:49
pittiTest glib/gdbus-threading.test failed: Child process killed by signal 509:49
Laneystill gdbus though09:49
Laneyso at least the same area09:49
pittisame on the next failure09:49
LaneyI think we know someone who speaks dbus quite well09:49
* Laney looks west09:49
Laneycould try to get it in an accessible instance09:50
pittihttps://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-xenial/xenial/s390x/g/glib2.0/20160118_105206@/log.gz is pcre, but I think that was due to pinning09:50
Laneypcre broke every arch09:50
Laneyand is fixed09:50
pittiLaney: can you ssh ubuntu@ ?09:50
pittiLaney: I mean firewall-wise (I didn't import your key yet)09:51
Laneypitti: yeah09:51
LaneyI know the password :)09:51
* Laney h4x0r09:51
pittiLaney: ah, ok -- I just did ssh-import-id lp:laney09:51
LaneyI've stolen this before to test stuff09:51
pitti"12345, the same combination that I use on my luggage!"09:52
pittiLaney: so I guess I'll just retry some more, and if I get bored I force it?09:52
pittinot really libselinux' fault anyway09:52
Laneywell... I'm going to upload a new upstream release to experimental today09:53
Laneyand sync that later or tomorrow09:53
pittibut nice, glib2.0 is the last s390x specific regression, I managed to fix all the other ones09:53
pittiall other regrssions happen on all arches09:54
Laneyadt-xenial-rva0ksiq  RUNNING  -     -       NO09:58
Laneyubuntu@aupkg01:~$ lxc-console -n adt-xenial-rva0ksiq09:58
Laneyadt-xenial-rva0ksiq is not running09:58
Laneythanks lxc!09:58
pittiLaney: is that s390x?09:59
pittiah yeah, that09:59
pitti$ sudo lxc-attach -n adt-xenial-rva0ksiq systemctl  status09:59
pittiLaney: ^ that clearly works09:59
pittiLaney: but I've never used lxc-console10:00
Laneypitti: ah right10:01
Laneylxc-start-ephemeral printed out that command for me to use10:01
Laneyoh right, sudo, /me fails10:01
pittiah, I finally know how to fix these eternal armhf hangs10:06
* Laney runs one of the tests in a loop10:06
pittiLaney: hah! lucky day! http://autopkgtest.ubuntu.com/packages/g/glib2.0/xenial/s390x/10:06
pittihttp://autopkgtest.ubuntu.com/packages/g/glib2.0/ -> IMPECCABLE!10:07
Laneyperfect package, no problems here10:07
* Laney coughs10:07
LaneyGLib-GIO:ERROR:/build/glib2.0-dQLvhN/glib2.0-2.47.4/./gio/tests/gdbus-proxy.c:826:fail_test: code should not be reached10:07
Laney# GLib-GIO:ERROR:/build/glib2.0-dQLvhN/glib2.0-2.47.4/./gio/tests/gdbus-proxy.c:826:fail_test: code should not be reached10:07
* Laney saves this state for a certain ARL10:08
pittioh, ARL!10:08
* pitti hugs desrt10:08
Laneylarsu: heeeeeeeeeeeey larsu10:14
Laneygtk css help required10:15
Laneylarsu: I was looking at webkit scrollbars (see devhelp on xenial for example)10:16
Laneyas far as I can tell it is drawing them itself instead of using gtk's ones10:16
Laneyhttps://sources.debian.net/src/webkit2gtk/2.10.4+dfsg1-1/Source/WebCore/platform/gtk/ScrollbarThemeGtk.cpp/ <- this is the file in question10:16
Laneybut there is some bug with drawing the background, it gets glitchy after you scroll10:17
Laneyhttps://paste.ubuntu.com/14581466/ adding those two rules makes it look somewhat normal (probably acceptable)10:18
Laneywondering if you can help suggest the right way to do it10:18
willcookehey andyrock10:45
larsuLaney: hm no clue. Let me check11:03
larsudoes it work with Adwaita?11:03
Laneyumm didn't actuallyt ry11:05
larsuit does11:06
Laneyyeah sort of11:06
larsusort of? Works fine for me11:06
Laneythey are obviously still custom scrollbars there though11:06
larsuah, true11:06
larsumaybe they just don't draw them in that case, but let ebkit handle them?11:06
Laneyyeah but still themed11:07
Laneyso I guess we can find out which rules are needed and put those in11:07
larsuhm but I can't find anything :/11:08
larsuLaney: they're not doing anything webkit specific, but setting backgrounds (like your patch)11:11
larsuyour patch makes it work for webkit windows, but breaks the style of other windows11:17
LaneyI know, it applies to too much11:17
Laneyjust wanted to see if it was fixable by styling at all11:18
LaneyI also don't know why you have to have the .scrollbar.slider rule since the same thing is in the theme already11:18
Laneybut there's no slider without it11:18
larsuI see it without it11:19
larsuor I misunderstand what I mean :)11:19
Laneyweird, I need both rules11:21
Laneylarsu: do you think you could look at this please? if you have time11:25
larsuyep, already doing it :)11:25
LaneyI have a load of stuff ready to move to webkit211:25
Laneyand kick the shitty old one out11:25
Laneykeep bumping into mcatanzaro on bugs too11:26
Laneyseems he is also doing this for fedora11:26
Sweet5harklibreoffice 5.1.0~rc2 finished building locally. now lets see about libreoffice-l10n11:44
larsuLaney: could you try it out please? https://code.launchpad.net/~larsu/ubuntu-themes/webkit-scrollbar12:06
Laneylarsu: is it less rounded than the regular scrollbars?12:21
Laneyotherwise it looks good to me12:22
Sweet5harkquestion: how do you get gcc to commit an internal compiler error?12:22
Sweet5hark(gcc6 that is)12:22
Sweet5harkanswer: LibreOffice, based on technology breaking your toolchain since 1985 ...12:23
larsuLaney: no change in the roundedness12:27
larsubut appears like it to me as well12:27
Laneymight be because they are wider12:28
Laneyso look more square12:28
larsuI wonder why12:29
larsushould be the same as expanded overlay scrollbars12:29
larsuLaney: ah, webkit seems to calculate slider width slightly differently. Good enough to keep like this?12:31
Laneylarsu: yeah if it's not fixable in the theme12:36
Laneyjdstrand: hi, would it be possible for someone to get to bug #1475021 quite soon please?12:58
ubot5bug 1475021 in gcab (Ubuntu) "[MIR] gcab" [High,Triaged] https://launchpad.net/bugs/147502112:58
Laneywe have appstream-glib in depwait for some time on this12:58
* Laney back soon13:00
ubot5Launchpad bug 1534085 in unity-control-center (Ubuntu) "System Settings legal text contains a 404 link and some inaccurate wording about online searches" [High,Fix released]13:21
willcookethanks Laney13:21
Laneywillcooke: woot14:08
desrtLaney, pitti: good morning14:16
desrteveryone else, good morning as well :)14:16
Laneyhi desrt!!!!14:18
desrtLaney: what's going on here?14:18
LaneyEVERYTHING (not much) is BROKEN (sometimes doesn't work)14:18
Laneysome gdbus tests seem not very good on s390x14:19
desrti noticed that assert is tied to a timeout...14:19
desrti don't suppose the test machine is super-slow...14:19
Laneyi think that it is the opposite14:20
Laneydo you have VPN?14:20
Laneyssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no ubuntu@
* desrt clicks a couple of buttons and lets NM do the work14:21
Laneythen screen -r14:22
desrtdoes ubuntu@ have a password? :)14:22
LaneyI imported your key14:22
Laneybut you could try to guess the password14:22
desrti'm in14:23
desrtmy yubikey seems to have trouble resuming from suspend...14:23
Laneyscreen window 0 and 1 show the two which have failed on the real instances lately14:23
Laneyand FYI you are inside an lxc container there14:23
jdstrandLaney: re 1475021 - I'll talk to tyhicks about it14:24
Laneythanks jdstrand, we'd like to get the ground prepared for uploading gnome-software quite soon14:24
Laneyso that the universe can explode with enough time for us to glue it back together14:24
larsumorning desrt!14:24
larsuhow goes?14:24
desrtwishing i had glasses14:25
larsucome to Berlin. Took them 1 day14:25
desrti went to a place here.  they did the test and sent me away.14:25
Laneyi showed rosie the picture of larsu with specs14:25
Laneyshe approved14:25
jdstrandLaney: ack14:25
larsuLaney: :)14:25
Laney"they look like mine"14:25
larsuLaney: say hi to her!14:25
Laneywill do!14:26
* Laney wonders how/what the ubucon people are doing14:27
Laneyguess it's still kind of early there14:27
jdstrandLaney: question for you-- just otoh, can you think of a reason why dlna wouldn't work on xenial in rhythmbox any more when it has the same upstream version as wily (and the Ubuntu delta didn't have any relevant changes), grilo is the same as wily and grilo-plugins only has one irrelevant change over wily?14:28
jdstrand(it the dleyna code in grilo that handles this afaik)14:29
desrtLaney: how do i get other screens into the container?14:30
Laneyjdstrand: hmm, nope, I tried this out a couple of weeks ago14:30
jdstrandLaney: the dlna server just doesn't show up any more (but daap still does)14:30
Laneydesrt: ctrl-a, <n>14:30
desrti know the screen part :p14:30
jdstrandLaney: it worked for you a couple of weeks ago?14:30
desrti'm missing the 'not in the container' part14:30
Laneyoh right, SSH to its IP14:30
Laneythought you might be a tmuxer or something14:31
Laneyjdstrand: well I could be confused about the difference between dlna and daap to be honest14:31
LaneyI shared my library between two rhythmboxen14:31
jdstrandI see14:31
=== ChrisTownsend1 is now known as ChrisTownsend
jdstrandI think rhythmbox exports as daap, iirc14:31
jdstrandexports/shares, whatever the terminology is14:32
Laneydoes it work with another client?14:32
jdstrandwell, yes, wily rhythmbox :)14:32
jdstrandthe minidlna server didn't change14:32
tjaaltoncodec search seems broken on xenial14:32
jdstrandI only upgraded to wily14:32
jdstrandI have both forked-daapd and minidlna exporting at home14:33
jdstrandand rhythmbox used to see both on wily. on xenial, it only sees daapd14:33
jdstrandtjaalton: it doesn't even see the server at all14:33
jdstrand(and it can play all the files from daapd)14:34
jdstrandso I don't think it is codec related14:34
Laneyok, I'll try to look14:34
Laneyis there a bug that you can assign to me?14:34
tjaaltonthis was just an observation on my local machine with totem trying to play bbb :)14:34
tjaaltonfresh install, can't find codecs to install14:34
jdstrandLaney: minidlna is really easy to setup. apt-get it, then adjust the config file for a dir of music14:34
tjaaltonthough I guess it's due to the software manager migration14:35
jdstrandLaney: I can. I'll work up a reproducer too. thanks14:35
LaneyI just want to be assigned something so I have less chance of forgetting :)14:36
desrtgdbus-proxy is failing on my machine this morning14:39
desrtdispite discheck having worked only two days ago14:39
desrtthat's neat.14:39
Laneytjaalton: looks like something happened to break sessioninstaller14:40
tjaaltonjust saw the trace14:44
desrtokay.  back to not failing.  great.14:58
* desrt made a mistake14:58
Laneytjaalton: https://paste.debian.net/366793/ ?15:05
tjaaltonLaney: hang on15:08
* Laney grips tight15:09
desrtLaney: gdb on this machine seems to be ... not great15:09
desrteven with debug symbols installed there are fairly big gaps in the backtraces15:10
tjaaltonLaney: nope :/15:11
Laneytjaalton: did you kill the daemon?15:12
Laneyit fixed it for me15:12
Laneyxnox: is gdb on s390x supposed to be working properly?15:12
tjaaltonLaney: correct, works15:12
xnoxLaney, yes....15:13
xnoxLaney, do you have bugs =)15:13
Laneymeet desrt15:13
Laneytjaalton: cool!15:13
LaneyI'll upload that then, thanks for reporting15:13
Laneyjust kicked a stair and now my toe is bleeding15:13
Laneyprobably should have kept the nail shorter :/15:14
* Laney goes to clean that up15:14
desrtxnox: hi.  seeing weird backtraces on a x390 machine :)15:14
xnoxdesrt, x390 ssl cert? or s390x machine?15:14
desrtuh.  the other one :)15:14
desrtisn't the ssl thing x509? :)15:14
tjaaltonLaney: thanks :)15:15
pittiLaney: ok, I think I should have finally stopped the worker log spam, sorry15:23
Laneypitti: pinning fun I assume15:23
pittiLaney: that, and unclear to me why it happened; but I built a new wily base image, it only happened for wily ..15:24
pittiLaney: and then I screwed up a fix from this morning which was supposed to fix the s390 tmpfails15:24
pitti(which didn't show in local testing)15:24
pittibut should be all good now15:24
pittiwell, I still need to fix the eternal armhf hangs, but at least I now know how15:24
pitti(and what happens)15:25
pittii. e. I can't fix the hung containers, but at least have adt-run ignore them insted of waiting forever15:25
desrtLaney: cannot reproduce on a fresh build >:|15:29
desrtLaney: can you tell me a bit more about the setup here?  these are installed tests running against a new library version, or what?15:29
Laneydesrt: build/install versions should be the same15:31
Laneythey are both 2.47.4-1 there15:35
andyrockif i do15:50
andyrockg_signal_emit (...., g_strdup(...))15:50
andyrockI'm supposed to g_free the string in the callback right?15:51
andyrockdesrt: ^^^15:52
desrtdon't do that :)15:52
andyrockor it's automatic?15:52
desrtg_signal_emit will fire the signal directly in place15:52
desrtso as long as the string in question will survive until g_signal_emit() returns, you don't need to dup it15:53
andyrockso I can free after g_signal_emit15:53
desrtin fact, depending on the signal, it may also dup the string for itself (and free it)15:53
desrtyes.  definitely.15:53
desrtbut maybe you don't need to dup it at all, is what i'm saying...15:54
desrtbut in any case: absolutely do not g_strdup() something you pass to g_signal_emit() and free() inside a handler.  that is absolutely broken.15:55
desrtand not just 'in theory' -- there is no guarantee that the string passed to the handler will have the same pointer value as the string passed to g_signal_emit() -- only that it will have the same value (as a string)15:55
andyrockok thanks! :)15:55
qenghoIf I'm merging a package from Debian, and I quote the patches that will remain in the package, should I also reference the Debian bugs with "closes: #nnnn"? That seems weird.16:31
LaneyYou don't have to, but it's nice to have them referred to somewhere16:32
Laneythe patch header is fine for that16:32
=== rickspencer3_ is now known as rickspencer3
desrtLaney: got a working theory17:13
desrtit's a stretch, though17:14
qenghodpm: Hi. Some translations team didn't respond to join requests. I advised a would-be contributor to message admins individually, wait, and then open a Question against Launchpad. Was that right?17:17
dpmqengho, it used to be, but nowadays no one really looks at those questions. The best thing might be to send an e-mail to ubuntu-translators@ubuntu(dot)com17:18
Laneydesrt: real bug or test bug?17:19
desrttest bug17:20
desrtlike usual17:20
desrtstill investigating...17:20
Laneyno need to waste all day on it if you have better things to be doing17:21
desrtLaney: actually, i've binned that first theory and am now tending toward 'actual bug'17:35
desrtthe proxy does a sequence of things in order to get the service to start up... first it tries to StartServiceByName... and if that fails then it does GetNameOwner in order to figure out if it already exists17:36
desrtthe problem is that maybe for some random reason (like a well-placed exec()) the service might pop into existence in between those two and the process may receive a NameOwnerChanged17:37
desrtin both cases, gdbus will try to GetAll properties on the service... which it seems to be doing twice17:37
desrtthat much, at least, is a bug in the proxy17:37
desrtso ya... one of those this-timing-issue-would-never-happen-in-the-real-world issues17:38
desrtbut it's a legit issue17:38
Laneyand I guess the test is making a new proxy async then execing the service17:40
desrtthe weird thing is that it should result in the async call finishing twice, as far as i can tell17:41
desrtand instead, it's finishing zero times17:41
desrtvery odd17:41
* desrt curses the gigantic ugly asyncinit state machine in here17:42
desrthilariously, this could just be an entirely unrelated bug because it also happens in the case where the test runs properly17:44
* desrt sighs17:44
* Laney hugs desrt 17:58
Laneyit's going to be okay17:58
* Laney goes off to roast a chicken17:58
=== lifeless_ is now known as lifeless
Mirvis there a bug for "lightdm doesn't show user background anymore and theme seems a bit resetted (domain name has grey background and date is black font etc)"18:13
MirvI have had it before 14.04 out of the blue and it happened again. I can't figure out how I fixed it last time, but I remember it was not straight-forward.18:13
Mirvlike uninstalling / resetting config files for lightdm + unity-greeter not enough18:13
=== xnox_ is now known as xnox
=== Tex is now known as Texou
MirvI meant, before on 14.04 and now again on 14.04. well, I'll eventually have time to find out the fix and at least document that somewhere.18:32
willcookenight night all19:01
=== ochosi_ is now known as ochosi
=== Guest6639 is now known as fredp
=== fredp is now known as Guest83051
attenterobert_ancell: hey, which branches of gnome-software should i be looking at?22:12
robert_ancellattente, wip/rancell/ubuntu-reviews-3-1822:12
robert_ancellattente, that's been merged into ubuntu-reviews22:13
attenterobert_ancell: ok22:14
attenterobert_ancell: i can't see ubuntu-ratings on the remote though22:14
robert_ancellattente, https://git.gnome.org/browse/gnome-software?h=wip%2Francell%2Freviews-3-1822:15
robert_ancellI'll delete the other branches22:15

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