=== RAOF is now known as tmpRAOF [07:08] good morning [07:32] good morning desktopers [07:33] re seb128 [07:33] lut didrocks :-) [08:06] bonjour didrocks et seb128 ! [08:07] lut pitti, ça va comment ? [08:08] bonjour pitti :) [08:08] seb128: je vais bien, merci ! alors, trop de bugs :) [08:18] seb128: hm, today's live image is broken -- something creates ~ubuntu/.config as user root and thus makes it not writable; causes the desktop to mostly fail [08:18] chown'ing it on VT1 and kill -9 -1 helps [08:20] pitti, shrug, I guess that's my fault :-/ [08:21] pitti, I added a snippet to casper that mkdir a subdir in there [08:22] let me have a look to fixing it [08:22] seb128: there's a .config/ubuntu-systmre-settings [08:23] seb128: with "wizard-has-run" [08:23] yeah that's it [08:23] seb128: thanks [08:25] sorry for the issue [08:25] also the desktop shouldn't fail to start because a dir is not writable :-/ [08:26] seb128: well, ~/.config is kinda important.. [08:26] well, you could also have a full disk [08:27] that shouldn't let you out in the cold === dupondje_ is now known as dupondje [08:45] morning [08:54] morning [08:55] hey willcooke mlankhorst [08:56] hey mlankhorst, willcooke [09:04] didrocks, pitti, http://paste.ubuntu.com/9913882/ looks good to you? [09:04] larsu: missing icons -> add to debian/not-in-humanity.txt in adwaita-icon-theme and upload it [09:04] good morning! [09:05] seb128: WFM; that's essentially what I did on the live system [09:05] Laney: morning! [09:05] Laney: but ... why?! [09:05] pitti, danke [09:05] seb128: OTOH, why do we need a "wizard-has-run" file on the live CD? [09:06] seb128: +1 [09:06] seb128: I got into the same trap 6 years ago, sorry, forgot to warn you about it :) [09:06] pitti, to not have the unity8 first run/configure your tz/password/wifi/etc on the live session [09:06] didrocks, :-) [09:06] Laney, larsu: I have already uploaded a-i-t with that icon added to landing-000 [09:06] larsu: we've had a split like this for years [09:06] * didrocks remembers to have broken the live during christmas period :) [09:06] mitya57: ok, why a silo? [09:06] hey larsu, wie gehts? [09:06] and uploading the fix on the 25 of december ;) [09:06] hey larsu [09:07] morning Laney, larsu! [09:07] bonjout seb128, didrocks! [09:07] *bonjour [09:07] I'm tired [09:07] Laney: I wanted to have all three fixes in one place to ask seb128 to test it :) [09:07] the idea is to keep cd space down by avoiding shipping icons in multiple themes [09:07] mitya57: cool, thanks [09:07] larsu, mitya57, no fix for the greeter no-password-prompt issue? [09:08] seb128: it needs fixes in THREE packages [09:08] seb128: already in trunk [09:08] larsu, can we land trunk? [09:08] unfortunately larsu's unity-greeter fix FTBFS [09:08] mitya57, we need changed to unity-greeter? [09:08] mitya57: really? robert_ancell reviewed it [09:08] from the team meeting yesterday I think to remember that robert_ancell is working on the build issues on vivid [09:09] seb128, https://code.launchpad.net/~larsu/unity-greeter/dont-set-spinner/+merge/247770 [09:09] but yeah, probably new vala or something [09:09] larsu: yes, and the bot agrees with me [09:10] robert_ancell summary at the meeting has "- Unity Greeter update for vivid (tests are failing on vivid, not on 14.04)" [09:10] mitya57: there's a mitya51 that is not you?! [09:10] * larsu is confused [09:11] there are lots of mityas [09:11] interesting. sorry about that. [09:12] hm, build works for me [09:12] no problem, I noticed the branch from IRC logs [09:12] and I doubt robert_ancell didn't test that [09:12] larsu, mitya57, hum, k, so if unity-greeter can't be build/uploaded, we might need to revert the theme change that broken it until u-g issue is resolved [09:12] larsu, test what? [09:12] did you try the testsuite? [09:12] larsu, u-g fails tests on vivid according to the meeting summary from robert_ancell [09:13] seb128: he approved the change, so I figured he'd have tested it [09:13] larsu, yeah, the change probably doesn't create regressions, still some u-g tests fail on vivid and that blocks upload to the distro until resolved [09:13] probably despite the already known test failure [09:13] larsu, which is orthogonal to your change but blocks its landing [09:14] ya [09:14] Laney: `make check` runs for me. I have to admit I didn't build the package though. Just from trunk [09:15] seb128: maybe I can just upload the fix with disabled tests? [09:15] seb128, mitya57: please don't revert the theme [09:15] I can give you a small patch that should make this work [09:16] mitya57, I would prefer not disabling tests, it might be that there is something wrong in vala in vivid and that the binary is going to be buggy in result [09:16] but also shows static images for proper spinners when they are not spinning [09:16] morning [09:16] larsu: small patch for -themes that doesn't require changes to greeter? [09:16] that would be ideal [09:16] is there anyone from the MIR team here? [09:16] pitti: I'm paying a high price in rebasing just to disable plymouth (as I did readd in multiple commits), ensuring that each commits can build, fixing typos, rebasing… and so on :p [09:17] oops, uploaded a ftbfsing a-l-m [09:17] mlankhorst: I am [09:19] mitya57: we need the change to the greeter - this patch would just trade bugs [09:19] didrocks: can you promote libepoxy in trusty? for xorg-server-lts-utopic. old bug at https://bugs.launchpad.net/ubuntu/+source/libepoxy/+bug/1342605 [09:19] Launchpad bug 1342605 in libepoxy (Ubuntu) "[MIR] libepoxy" [High,Fix released] [09:19] mlankhorst: the MIR seems to have been already approved, right? So you just need an archive admin, not a MIR team member (some MIR team member aren't aa) [09:20] larsu: the adwaita-icon-theme change is needed in any case, right? [09:20] ok I just wasn't sure because this is an older version than the one from utopic that was approved [09:21] mlankhorst: did you upload xorg already? [09:21] yeah [09:21] mlankhorst: you should get it into -proposed, and then, I can (with my AA hat) promote the deps [09:21] mitya57: yes [09:21] mlankhorst: not accepted to -proposed as far as I can see [09:21] https://launchpad.net/ubuntu/+source/xorg-server [09:21] I wish we could fix the greeter tests instead though [09:21] didrocks: https://launchpad.net/ubuntu/trusty/+source/xorg-server-lts-utopic ? [09:22] let's look at the tests ;-) [09:22] yes, please [09:22] mlankhorst: waow, the summary page doesn't show it! [09:22] https://launchpad.net/ubuntu/+source/xorg-server-lts-utopic/2:1.16.0-1ubuntu1.2~trusty1 [09:22] mlankhorst: yeah, seeing here [09:22] oke, good :-) [09:22] mlankhorst: let me do a quick sanity check in term of deps [09:24] mlankhorst: ok, no deps change, promoting [09:24] Laney, larsu, unity-greeter fails to build for me as well [09:24] mlankhorst: I'm quite unsure how this will work btw as it's not in -update, let's see [09:25] no idea either [09:26] seb128: tbh, I think we should upload the fix... [09:26] larsu, which one? [09:26] seb128: clearly the test is not related to my branch [09:26] larsu, well, the package is going to fail to build [09:26] or what do you mean? [09:26] you suggest disabling the testsuit? [09:26] yes [09:26] I don't like that, there might be real problems due to the new vala or something [09:26] is the version we have in vivid now failing to build as well? [09:27] we should understand the errors [09:27] yes [09:27] mlankhorst: yeah, that doesn't work [09:29] hm iirc some package in precise got promoted for similar reasons, but can't remember if that was the case or not [09:29] larsu, the tests have warnings like [09:29] (./unity-greeter-test:14152): Gtk-WARNING **: Could not find the icon 'image-missing-ltr'. The 'hicolor' theme [09:29] was not found either, perhaps you need to install it. [09:29] that should not be related [09:29] should be fixed anyway... [09:29] mlankhorst: pinged colin about it [09:29] larsu, yeah, not sure what we are arguing on there [09:30] see #ubuntu-release [09:30] oke [09:30] larsu, robert_ancell wrote in his weekly summary that it's working on fixing the tests, but that didn't happen yet [09:30] larsu, if it takes a week we don't want to have the password prompt missing on the greeter for a week [09:30] I can drop an email to robert and see tomorrow where he is with those issues [09:31] seb128: wait for a moment. I'm still trying to reproduce [09:31] seb128: do you see the same problem when building in source? [09:31] * larsu doesn't get any of those warnings [09:31] no, only with debuild [09:33] I guess another env issue [09:37] larsu, unset XDG_DATA_DIRS make check [09:37] that reproduces it [09:37] the env in which the tests are run doesn't have the XDG_ variables it seems [09:37] seb128: thanks!! [09:38] larsu, yw! [09:38] they should definitely run with those set [09:38] shouldn't crash without them, though [09:39] yeah, not sure what change the env in the packaging tools === jhenke_ is now known as jhenke [09:41] the test env should set them afaic [09:41] *afaik [09:42] bug in the tests then? [09:44] bug in the system whereever the tests are running I would say [09:45] Hm, I seem to get a message about hicolor not found even if I install hicolor-icon-theme [09:46] see previous chat [09:46] mitya57, it's because XDG_DATA_DIRS is not set [09:46] XDG spec says you should cater for them being unset or empty [09:46] this is a bug in gtk even [09:46] why? [09:46] the test runner sets XDG_DATA_DIRS [09:47] Laney: sure? I get the exact same error when unsetting it [09:47] look at tests/test.vala [09:47] * Laney builds [09:48] so the problem in gtk is this: it doesn't find an icon and falls back to image-missing-ltr, but doesn't check whether _that_ is there [09:53] hm, that icon is shipped with gtk [09:54] larsu: do you want a stacktrace? http://paste.ubuntu.com/9914563/ [09:55] thanks, but I've already figured it out [09:56] seb128: so ... these tests can be fixed by a gtk upload :) [09:57] great! [09:57] what's the issue? [09:57] we don't ship the icon in our binary? [09:57] e.g .install to update? [09:57] seb128: gtk doesn't. It has the icon in tree, but fails to include it as a resource (a simple Makefile.am change) [09:57] oh ok [09:58] seb128: do we have the time for this? Or do you want to revert? [09:58] * larsu needs a bit to test everything [09:58] larsu, let's move forward [09:58] you can also fix the test to set up XDG_DATA_DIRS properly and install an icon theme [09:58] Laney, it's not set properly? [09:59] I though you said it was [09:59] well it doesn't add the system directories if the variable isn't there [09:59] k, which is the case there [09:59] Laney: all packages have to do this? [10:00] It sets the variable up itself to point to a private directory [10:01] so it ends up being called like XDG_DATA_DIRS=/.../unity-greeter/usr/share/ ./testsuite [10:01] instead of XDG_DATA_DIRS=/usr/local/share/:/usr/share/:/.../unity-greeter/usr/share/ ./testsuite [10:02] Laney: it appends the system dirs, though [10:02] where? [10:03] also couldn't this set GSETTINGS_SCHEMA_DIR instead? [10:04] hm? What does that have to do with it? [10:05] test.c:758 [10:05] didrocks: err, I was sure I read "Exit statuses" somewhere, sorry about that [10:05] didrocks: hence my recommendation to use "status" or "states" -- weird brain-o! [10:05] pitti: ah, I thought you wanted to remove the "Exit" [10:05] pitti: thanks again for all the reviews! [10:06] didrocks: I'm currently knee-deep in debugging two other things and have two meetings in the afternoon [10:06] didrocks: but I suppose there's not much further stuff that I'll find by myself, so feel free to post upstream! [10:07] larsu: one sec [10:07] pitti: ok, doing some final testing and preparing the email! good luck for your debugging :) [10:07] didrocks: thanks, and good luck with the upstreaming! [10:07] thx! [10:08] didrocks: lennart is currently on a "catch up with ML and patches" quad-damage spree, so good time :) [10:08] heh, indeed! [10:09] mlankhorst, did you see my message last night about input in Mir? [10:10] yeah [10:10] good news [10:10] larsu: http://paste.ubuntu.com/9914739/ [10:10] mlankhorst, is it? I was going to ask - does it look like we're getting closer to what we need? [10:10] I'll take "good news" to be a yes :) [10:11] seb128, I opened this too: https://bugs.launchpad.net/mir/+bug/1415161 [10:11] well it's not what you think it is, but it's useful for me [10:11] Launchpad bug 1415161 in Unity 8 "U8 application windows don't remember their location or size when switching between Windowed->Staged->Windowed" [Undecided,New] [10:11] mlankhorst, progress LD# [10:11] :D [10:11] gives: http://paste.ubuntu.com/9914743/ [10:11] Laney: that might be necessary, but doesn't fix the issue [10:11] seb128, I dont think it will be useful for pure desktop mode necessarily [10:11] afaict it's required to get rid of the separate mir threads, X is single threaded [10:11] Laney: and I wonder why it's compiling the schema itself anyway... [10:12] Laney: so that we can run the test in-tree? [10:12] I guess [10:12] but it doesn't mean input support is suddenly better [10:12] from what I can tell at least [10:12] mlankhorst, ohh, ok. [10:12] sad face [10:13] it fixes the *build*, and means that the test doesn't mess up XDG_DATA_DIRS (and presumably runs with right icons now) [10:14] but you're probably right about there being a gtk bug too [10:14] Laney: I'm pretty sure it doesn't fix the build [10:14] ... [10:14] I tried it [10:14] 28/01 10:11:10 gives: http://paste.ubuntu.com/9914743/ [10:15] oh, you added ubuntu-mono as well [10:15] fair enough, sorry [10:15] ya [10:15] * larsu focussed on the GSETTINGS_SCHEMA_DIR [10:16] Laney: thanks a lot. That makes things easier. Want me to just add this patch to my branch? [10:17] Laney: isn't humanity-icon-theme enough? [10:17] dunno [10:18] larsu: sure [10:24] Laney: done, thanks again [10:24] seb128: ^ [10:26] ty [10:26] mitya57: you want to re-upload to landing-000? [10:27] Laney: I can (though in this case it will be -0ubuntu3) [10:27] yeah, I don't know why this native package has a non-native version number [10:28] doesn't really matter [10:29] bonjours a tous! [10:30] pitti: I see the libreoffice autopkgtests for are red -- but the logs show a happy "junit-subsequentcheck PASS" on both i386 and amd64. Any hints? [10:31] junit-subsequentcheck FAIL non-zero exit status 1 [10:31] Laney: wait a minute [10:31] patching file solenv/gbuild/JunitTest.mk [10:31] Hunk #1 FAILED at 57. [10:31] 1 out of 1 hunk FAILED -- saving rejects to file solenv/gbuild/JunitTest.mk.rej [10:31] Sweet5hark: ^ probably due to that? ^ [10:35] * Laney freezes [10:36] * mitya57 wonders why dpkg refuses to build a .dsc [10:36] willcooke, ok, do they want us to file bugs about wm issues? I assumed that it was not very useful to do so yet, since they basically did 10% of the work needed, so you could file like a list of an hundred "bugs" [10:36] seb128, I saw a feature I wanted so I logged a bug :D [10:36] willcooke, sure, I'm just asking if we should spam them [10:36] pitti: where did you find that? so far I only found "hunk succeeded ..." at https://jenkins.qa.ubuntu.com/view/vivid/view/AutoPkgTest/job/vivid-adt-libreoffice/ARCH=amd64,label=adt/lastSuccessfulBuild/artifact/results/log [10:37] willcooke, like I can open some 30 of those todays if you want to track every detail of wm [10:37] we should maybe talk to them to know if they think it's useful there [10:37] seb128, I'll put it on the agenda for next week.... do they want feature requests from us. I think we should log them, since "we" have 10 years of knowing about a desktop WM [10:37] or if we should let them get it a better working state first [10:38] willcooke, right, some of them do as well, they used to maintain compiz :-) [10:38] fair [10:38] :) [10:39] seb128: "suffer from" you meant? [10:39] :D [10:39] Laney: uploaded to 000 [10:40] didrocks, that too indeed! [10:40] groovy [10:40] larsu, sorry, do you have something ready/that needs sponsoring? [10:40] seb128: that's what mitya57 just uploaded [10:40] seb128: I think mitya57 is on it [10:40] great [10:40] I think we can get the gtk fix via upstream now [10:41] right [10:41] assuming larsu still thinks it's worth it [10:41] Laney: definitely needed [10:41] * larsu is hacking Makefile.am right now [10:41] \m/ [10:41] * darkxst thinks mitya57 needs an ambulance ;) atleast thats what you get in Aus when you dial 000! [10:42] seb128, if your looking for things to sponsor ;) my gnome-session wayland split needs sponsoring ;) [10:42] darkxst, hey, I can try to have look today [10:43] seb128, thanks [10:43] darkxst: :D [10:43] darkxst: did bigon get back to you? [10:43] (oh hi bigon, didn't know you were here!) [10:44] we can pay him in beer over the week-end otherwise [10:44] Laney, no [10:44] Laney: oh, it again ftbfs [10:44] wtf [10:44] can it really be? [10:45] https://launchpadlibrarian.net/196018040/buildlog_ubuntu-vivid-ppc64el.unity-greeter_15.04.2-0ubuntu3_FAILEDTOBUILD.txt.gz [10:45] Sweet5hark: it's the latest log -- https://jenkins.qa.ubuntu.com/job/vivid-adt-libreoffice/122/ARCH=i386,label=adt/console (at the bottom) [10:45] * mitya57 tries to understand what happened [10:45] unmet build deps, what even? [10:45] Laney, my fault probably [10:46] darkxst: hey [10:46] bigon, hi [10:47] you should better ask to somebody else in the debian GNOME team, I'm quite unavailable ATM [10:47] Laney: I don't want to do a -0ubuntu4, so maybe I'll upload it directly to archive with -0ubuntu2? [10:47] hang on [10:48] that would be fine if you want to [10:48] (didrocks: I hope I'll be able to attend fosdem, that's not 100% sure ATM :/) [10:48] pitti: doh! thanks, I was confused by jenkins showing "last successful artifacts" were I expected "last artifacts" .... [10:49] bigon: argh, keep us posted! === Chipaca` is now known as Chipaca [10:49] bigon: and good luck :) [10:49] Laney: in fact, we should convince robert_ancell to select a better format than 3.0 native with revision number [10:50] All my tools refuse to work with that :) [10:50] AFAIK you just have to say 'y' to debuild -S [10:50] but indeed it makes them complain at you [10:50] bigon, ok I will keep pinging! [10:50] still don't see the problem, btw [10:52] oh right, the dsc doesn't have it [10:52] wonder how you managed that [10:55] waiting for the last test build to finish, will be really soon [10:55] can you improve my changelog message? [10:57] "* Set GSETTINGS_SCHEMA_DIR - this environment variable can be used instead of XDG_DATA_DIRS to cause gsettings to use a different directory for loading schemas." [10:57] "* Build-Depend on ubuntu-mono - the above change fixes a hidden bug which caused a fallback icon theme to be used. We can now use our proper one." [10:57] or something [10:57] mitya57: ^ [10:58] Laney: too late... [10:59] ho hum :) [11:00] Laney: actually I wonder how much sense does it make to build-depend on ubuntu-mono and not ubuntu-settings [11:01] How does Gtk know that our theme is ubuntu-mono? [11:01] /etc/gtk-3.0/settings.ini [11:02] Nice, didn't know about that [11:02] * mitya57 notices that gtk-fallback-icon-theme is still gnome, not Adwaita [11:03] Laney: oh wow, it still looks at that? [11:03] Or maybe I just need to upgrade gtk [11:03] I think so ... [11:03] at least I munged it in jhbuild the other day when testing reftests and it seemed to respect it [11:04] interesting [11:04] mitya57: actually that fallback one is ignored according to the documentation [11:04] ok [11:04] Now that greeter and a-i-t are in archive, should I land ubuntu-themes as well? [11:05] larsu: ah this is the fallback when there's no other xsettings manager [11:05] ah, right [11:13] larsu: can you please copy the changelog entry from what I uploaded to your MP? [11:14] I can just fix it up when pushing [11:15] thanks Laney [11:15] ok [11:16] themes are there, sponsored for larsu [11:40] popey, next sprint? https://www.kickstarter.com/projects/elanlee/exploding-kittens :D [11:43] yeah, I backed it :) === dgadomski_ is now known as dgadomski [12:05] guys [12:05] since yesterday on vivid [12:06] tsdgeos: we're on it. [12:06] lightdm is missing the text fields [12:06] i can still type the password [12:06] fix is in the queue [12:06] larsu: stop reading my mind, not funny :D [12:06] tsdgeos: ;) [12:08] \o/ === MacSlow is now known as MacSlow|lunch [12:26] note to self: try to run the test suite of the release series you want to test next time. [12:26] amazing that it did work that well ... [12:40] Sweet5hark, hum, libreoffice migration is blocked on libcmis autopkgtests issues that look like to be errors in the test machinery [12:41] hum, not on this one https://jenkins.qa.ubuntu.com/job/vivid-adt-libreoffice/lastBuild/ARCH=i386,label=adt/console [12:41] adt-run [17:12:41]: test junit-subsequentcheck: - - - - - - - - - - results - - - - - - - - - - [12:41] junit-subsequentcheck FAIL non-zero exit status 1 [12:41] Sweet5hark, can you look at that? [12:43] seb128: see backlog [12:43] Sweet5hark, oh, I didn't connect what you wrote to the issue [12:44] you said " amazing that it did work that well ..." [12:45] larsu, do you think you could nag somebody about https://bugzilla.gnome.org/show_bug.cgi?id=738625 ? [12:45] Gnome bug 738625 in general "Add = and = keybindings for zooming in" [Normal,Unconfirmed] [12:46] seb128: yeah, I locally tested the 4.3 testsuite against 4.4 ... and that worked, which is pretty amazing. However we need 4.4 tests running against 4.4. I already have a http://people.canonical.com/~bjoern/vivid/4.4.0/libreoffice_4.4.0~rc2-0ubuntu3_source.changes but hang on, Im still testing if that really fixes the issue. [12:50] seb128: yeah, I locally tested the 4.3 testsuite against 4.4 ... and that worked, which is pretty amazing. However we need 4.4 tests running against 4.4. I already have a http://people.canonical.com/~bjoern/vivid/4.4.0/libreoffice_4.4.0~rc2-0ubuntu3_source.changes but hang on, Im still testing if that really fixes the issue. (resend, got disconnected) [13:04] * Laney stabs autotools [13:05] FUNCTION ([args...], FOO) → possibly undefined macro FOO [13:05] what it really means is that I shouldn't have put a space after FUNCTION [13:05] what kind of error is that [13:06] Laney: it can't know, because macros without arguments exist [13:07] why didn't I get an error which talked about FUNCTION? [13:07] that's where the actual error was [13:08] because you successfully defined a macro called FUNCTION [13:08] every occurrence of FUNCTION in your source gets replaced with ([args...], FOO) [13:08] and then it doesn't know about FOO [13:15] * Sweet5hark1 joins in stabbing autotools. for whatever reason. it deserves. === MacSlow|lunch is now known as MacSlow [13:41] seb128 / didrocks: could you quickly merge lp:~laney/activity-log-manager/merge-archive-uploads into trunk for me please? [13:42] Laney: sure, looking [13:42] Laney: done [13:43] Laney, want access to the corresponding team? [13:43] didrocks, thanks [13:45] seb128: if you want ;-) [13:46] oh, in fact I'm not admin, can't do that [13:47] oh well [13:47] thx for looking [13:47] thanks didrocks [13:52] yw [14:18] db3xj [14:19] whops ignore that [14:19] Laney: what does those "apply upstream" means? they are in ubuntu-control-center now? [14:19] It means I applied a distro patch upstream [14:20] Laney: upstream isn't lp:activity-log-manager? [14:20] ah, you kept that in the same repo [14:20] just applied it inline [14:20] it never was in the repo, I just uploaded it [14:21] Laney: bzr repo [14:21] correct [14:21] ok, I'll probably need to touch the whoopsie part, hence the question [14:22] (1) apt-get source a-l-m (2) hack-hack-hack (3) dput (4) bzr branch lp:a-l-m (5) patch -p1 < ... (6) bzr commit (7) bzr push [14:22] yeah, sounds good [14:22] maybe you could just release what's in bzr though [14:22] Laney: yeah, can't :/ [14:22] as you have access there [14:22] I think robert_ancell just applied all the distro patches [14:31] seb128: eh, revitalizing the autopkgtests on 4.4 might take a while ... :/ [14:36] Laney: I've backported the image-missing patch to our package. Not sure if it's worth it tbh, but maybe we'll see similar issues in other tests [14:36] https://code.launchpad.net/~larsu/gtk/backport-image-missing-patch [14:36] Laney: do you prefer a MR? [14:37] larsu: nah, that's fine, ta [14:37] will you push it to 3.14 when acked? [14:38] not sure [14:38] probably a good idea... [14:38] think so [14:39] mclasen does to. Pushing [14:39] I guess you can ignore my branch then and wait for the next release [14:40] ok [14:40] will keep it in mind if we see anything similar [14:50] we can also stack in the vcs for the next upload [14:51] not sure how often gtk upstream is still going to release on that serie [14:53] seb128: there was one 17 days ago [14:53] I guess we could just ask mclasen to make one if we need it [14:55] we don't, no hurry [14:55] I guess there is going to be one at some point still this cycle [14:56] * larsu nods [14:56] if not, we'll ask then [14:57] wfm [15:23] pitti: quite unsure about bug #1411140 TBH, the failure that I can reproduce seems to be that we can't unmount /etc/machine-id as if it wasn't a mount point [15:23] bug 1411140 in systemd (Ubuntu) "systemd-machine-id-commit.service fails on live system" [Low,Triaged] https://launchpad.net/bugs/1411140 [15:23] pitti: but the job start condition is that /etc/machine-id is a mount point… [15:24] pitti: and if I set that live up (after boot) with / ro, machine-id-setup -> / rw [15:24] and run machine-id-commit, no error [15:24] didrocks: oh nice, you can use send() on stream sockets? I wasn't aware of that (just reading your new patch) [15:24] pitti: yeah ;) [15:25] didrocks: might be some weird interaction with overlayfs? [15:25] pitti: funny that part of the comments are on the old code I just moved between files though :p [15:25] pitti: yeah, that's my guess :/ Not really sure how to handle this though [15:25] like how to test apart from copying shell and break in casper bottom? [15:25] didrocks: not really urgent IMHO; it's just cosmetical, we can handle that after FF [15:26] pitti: yeah, I would love to understand it though :p [15:27] pitti: oh, seems that if I stop it and restart, it's still starting even if /etc/machine-id isn't a mount point [15:28] * didrocks smells that Jan 15 08:08:47 ubuntu systemd[1]: Starting Commit a transient machine-id on disk... [15:28] ConditionPathIsMountPoint is ignore… [15:28] ignored* [15:29] pitti: right, seems it's overlayfs messing with ConditionPathIsMountPoint [15:35] didrocks: yeah, doesn't surprise me too much; I figure systemd still sees the underlay? [15:37] pitti: yeah, can be [15:50] autopkgtests are running now for some time ... [15:50] * Sweet5hark1 hopes for the best [16:02] ... and finished. [16:03] seb128: http://people.canonical.com/~bjoern/vivid/4.4.0/libreoffice_4.4.0~rc2-0ubuntu3_source.changes should fix the autopkgtests. Please consider uploading. The diff is small (only in ./debian/tests so should not risk any regressions outside of autopkgtests) [16:44] Sweet5hark1, was that a real issue or a buggy test? [16:44] if it's only a buggy test maybe pitti or Laney can help overwrite the autopkgtest result to get the current build to migrate to vivid? [16:45] seb128: the patching of the test to only run the subsequenttest (and not a full build) was outdated. [16:45] seb128: so no failing test and thus no change to production code [16:46] I can help unblocking packages which are blocked on *other* package's failing tests (although the release team would frown), but as I'm not ~ubuntu-archive I can't add the official "force-badtest" stamp to promote a package with broken tests [16:47] I'm fine with having that added for current LO to avoid another long upload/build just for that, though [16:47] Wouldn't it be best to know that the tests pass? [16:47] sure :) [16:48] Laney: sure, but in this case you are squatting an armhf builder for ~a day to rebuild stuff with no change at all ... [16:49] well, half a day [16:52] It's not no change though, the change is to make the tests actually get run? [16:53] If you mean no change to LO itself, well yes, but we can't just change debian/tests on it sown [16:54] Can override it if you really prefer though [16:57] Laney: the diff is changelog and a patch in ./debian/tests/patches. No change to ./debian/rules, ./debian/patches ./debian/control, so it should produce the same binaries (modulo changes in dependencies in the meantime). [16:58] I get it - the point would be to get a successful test run on jenkins.qa.u.c [16:59] Laney: FWIW, this is rc2 and rc3 was tagged upstream yesterday and will be released as final tommorrow. So this build wont stay around forever. [17:02] Laney: anyway, I mostly dont care how to move this forward: either by handwaving on jenkins or by uploading a 0ubuntu3. I care though that this moves forward as there are more changes to be done (rc3 from upstream, mergedlibs, some MIRs and dep changes etc.) and I dont want to stuff even more in this first-upload-of-a-major-release-series. incremental changes are much better. [17:03] Sweet5hark1: OK I added the skip. For this version only, so make sure the next one works. :) [17:03] http://packaging.ubuntu.com/html/auto-pkg-test.html#executing-the-test ← that's how I run these tests [17:04] Laney: yeah, I know. I ran the test locally. [17:08] Laney: Before -0ubuntu2 I ran the tests too, but I fumbled by running the tests from LibreOffice 4.3.3 against 4.4.0 ... which amazingly worked, but made me miss that I needed to refresh the patch against 4.4 (which is what I did for -0ubuntu3 and then ran the right testsuite for a change) [17:08] seb128: I have a fix for bug 1415104 :) [17:08] bug 1415104 in systemd (Ubuntu) "logind flags wrong session as active" [Medium,In progress] https://launchpad.net/bugs/1415104 [17:51] pitti: can I try out systemd on vivid yet? Would like to help amigadave a bit on gnome-journal, but I have nothing in the journal on my system :) [18:16] pitti, great! [18:33] * willcooke -> EOD o/ [18:56] Laney: appdata-util and appstream-util should probably conflict each other [18:56] * larsu just got an error about both trying to install the same files [20:00] Laney, nice, you found what was breaking the u-g build === Zdra is now known as xclaesse === Sweet5hark1 is now known as Sweetshark