[03:50] Good morning [04:20] mornin' [04:41] hi [05:46] hey hikiko, how are you? [05:46] jbicha: do you need bug 1616198 sponsored or can/do you want to sync yourself? [05:46] bug 1616198 in gnome-desktop-testing (Ubuntu) "FFe: Sync gnome-desktop-testing 2016.1-1 (universe) from Debian unstable (main)" [Wishlist,Confirmed] https://launchpad.net/bugs/1616198 [05:47] Hi pitti ! I'm fine :-) you? [05:47] hikiko: woke up too early, but fine too, thanks! [05:49] pitti: I'm a motu so I can sync it, I just didn't want the release team too upset with me :) [05:49] Happens :-) you ll finish work earlier now!! [06:02] hikiko: heh -- work never finishes, you just stop at convenient times :) [06:07] hahaha, yeah, that convenient time will come earlier today! [08:01] Morning all [08:03] yoooo [08:06] moin [08:06] good morning desktopers! [08:06] * Sweet5hark had installed cups debug symbols and started a libreoffice build without disabling cups in a sbuild [08:07] this went as expected: it ran towards the tests overnight and then hung. [08:07] jbicha: gnome-terminal> nice [08:07] hi Sweet5hark & seb128 [08:08] and then sbuild killed the whole mumbo jumbo after 150 minutes before inactivity -- so when I came back to look at it there was nothing left .... [08:09] s/inactivity// <- me needs coffein. [08:09] hey Sweet5hark, no luck, you can play again though! [08:09] hey Laney, how are you? [08:09] seb128: yep [08:09] you could probably hax0r sbuild to not do that [08:09] seb128: good! [08:10] the sun is here [08:11] Morning folks [08:11] you? [08:11] hey Trevinho, wie gehts? [08:11] Hey Laney.. [08:11] i saw the italy news :( [08:12] hey Trevinho, how are you? hope you are not/don't have familly/friends impacted by the earthquake [08:12] Shocked by that... 😢 [08:12] but yeah :-( [08:12] No, as far I know... [08:12] But still... [08:12] yeah :-/ [08:13] Laney, summer really kicked in here, it's already 27°C and should get to 34°C today [08:13] And I didn't notice that either, although people around me waked up by the vibrations [08:13] Laney: well, yeah. Mostly Im just grumpy at myself for not considering that .... [08:14] morning al [08:14] l [08:15] hey willcooke, feeling better? [08:15] seb128, yes thanks! [08:20] happy birthday hikiko! [08:21] oh! hikiko, happy birthday! [08:22] hello folks, good morning! [08:22] hikiko: happy birthday from me too! [08:23] hey pitti, good "evening"? ;-) [08:23] * seb128 saw on the irclog that you start at like 3utc [08:23] seb128: hehe, en effet -- j me suis levé à 5h15 aujourd'hui :) [08:24] bien, en fait j'ai commencé travailler dans le lit :) [08:24] thanks willcooke seb128 and pitti :D [08:25] * hikiko brings the beer and the cake :) [08:25] \o/ [08:25] * pitti adds some ice cream [08:25] :D [08:31] Oh happy birthday hikiko :) [08:32] thank you duflu! [08:32] My age is now a power of 2 :p [08:32] * hikiko feels old :) [08:35] hikiko: You're young enough to make me feel old [08:35] hikiko is 16? [08:35] happy birthday! [08:36] Indeed [08:37] lol Laney [08:37] χαρούμενα γενέθλια [08:37] hehe :) ευχαριστώ Laney ! [08:38] it's funny that there is a language made up of mathematical symbols [08:39] LOL [09:05] morning [09:05] hey andyrock, how are you? [09:07] a bit shoked for what happend but ok :/ [09:09] hey andyrock [09:10] and happy BD hikiko! [10:31] hey Trevinho, thanks! === hikiko is now known as hikiko|ln [10:32] :) [10:42] Laney: so, to make "LibreOffice on yakkety" the full yolo edition .... can you make it skip the autopkgtests that are failing on missing deps (not a test failure suggesting trouble with the binaries per se)? [10:42] seb128, jbicha: ^^ [10:43] <- not proud === hikiko|ln is now known as hikiko [11:10] Sweet5hark: which tests? [11:10] http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#libreoffice [11:17] bug #1349859? [11:17] bug 1349859 in liborcus (Ubuntu) "[MIR] libixion (b-d of liborcus)" [High,Confirmed] https://launchpad.net/bugs/1349859 [11:17] it was in main so I guess just need to be repromoted? [11:17] I can do that [11:18] I guess so, and then we should get test results [11:50] tedg: hey, did you find some time to update that systemd branch then? [12:16] hm. On 16.04 - after a few days of uptime, I'm unable to launch any graphical apps, everything gives me "Maximum number of clients reached" and i have to go round killing my GUI apps, or end my session or reboot. Anyone else seen this? [12:24] Hey, is there a metapackage that will always depend on the latest LTS HWE stack? We have to upgrade a bunch of machines and it would be nice not to have to do this by hand [12:41] Trevinho, hey :), any plans to provide a new bamf 0.5.2/3 tarball? [12:41] ricotz: hey [12:41] ricotz: well, nothing is planned right now... If needed i could do one [12:42] Trevinho, it has been over a year already since 0.5.1, so it might be worth to do one if no merges are pending [12:50] popey, hmm, no. LP has a few bugs of similar nature though, but most are quite old [12:52] it's pretty infuriating, has been happening for weeks now. but obv only happens after some days of uptime and apps left open. [12:52] i only notice once I click an icon and nothing happens [12:53] popey, long shot here (based on googling) - do you use lastpass? [12:53] i do [12:53] saw that. [12:54] popey, I'll leave my computer on for the rest of the week at least and see if I can reproduce it. I'm using Chrome but not lastpass [12:55] but like you say, they're super old [12:56] My gut feeling is that its an app not cleaning up after itself rather than a regression in, say, compiz. But lets see if I can reproduce it.... [12:56] could lightdm have a hand in here as well maybe.. humm [13:01] popey, try to "for x in `ps -eF| awk '{ print $2 }'`;do echo `ls /proc/$x/fd 2> /dev/null | wc -l` $x `cat /proc/$x/cmdline 2> /dev/null`;done | sort -n -r | head -n 10" ? [13:01] in case that's a fd leak [13:07] seb128: ooh, that's neat [13:08] chrome 540, dropix 127, atom 102, dbus-daemon 85, pulaseaudio 59 [13:08] *dropbox [13:08] so yeah, app problem it seems, thanks for that script [13:15] popey, yw! [13:21] Sweet5hark, Laney, k, now libreoffice has grumpy autopkg https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-yakkety/yakkety/amd64/libr/libreoffice/20160824_120855@/log.gz [13:28] seb128: *sigh* [13:28] thanks seb128 (re: pope_y 's problem) [13:28] Sweet5hark, do you see the issue in the log? [13:28] willcooke, yw! [13:29] seb128: please ignore for now. rene added testing for the sdk in libreoffice 5.2, we didnt test that before. I'd need to triage why they dont work. [13:30] Sweet5hark, ignore what? [13:30] I can't skip those, need to talk to p_itti or L_aney if you want to get it to migrate anyway [13:30] especially that there is a whatevermilestoneoftheweek blocker in place [13:31] seb128: ignore the failing autopkgtest, yes. <- Laney [13:31] * Sweet5hark grumbles: yeah, this all should have happened weeks ago., [13:38] ... or ^^ pitti: please skip the autopkgtests for libreoffice yakkety. [13:39] Sweet5hark: is that upload for beta-1? [13:39] Sweet5hark: i. e. super-urgent? [13:39] (beta freeze until tomorrow evening) [13:40] urgh, timed out? is that something which can be ignored safely? i. e. do you understand why it happens? [13:43] pitti: everything is horrible right now: we have libreoffice-l10n 5.2.0 promoted to yakkety, but only libreoffice 5.1.x in as libreoffice 5.2.0 is stuck in yakkety-proposed. I realize this is all way too late and undertested and it sucks $PFEEP, but yeah, Id say it is urgent. [13:44] that doesn't answer the above questions, though :) ("for beta1?", "can be ignored safely?" "do you understand what happens?") [13:45] pitti: the particular test is testing the horrible libreoffice-sdk that two people in the world actually use, not the usual libreoffice autopkgtests (which havent run yet). rene added this at debian and apparently it doesnt work for us (and as it was new likely never was). [13:47] pitti: so: for beta1: yes "can be ignored safely": this is not a regression, but a new test from debian that apparently doesnt work on our infra (and probably never was). normal autopkgtest where not run because of this, so we dont know if something serious would fail there too. I did a very basic startup test of the binaries in -proposed though on a VM. [13:48] pitti: "do you understand what happens": an test for the SDK added by Debian doesnt work for us. Dont know why yet, it likely never worked for us. [13:48] Sweet5hark: ok; I'll hint the package, but allowing it into beta1 needs to be done by the beta-1 RM (coordination with image rebuilds) [13:49] * Sweet5hark really really hates all this. [13:51] Sweet5hark: if you aren't sure, we could alternatively revert libo-l10n to 5.1; and you should add versioned deps to the -l10n packages to make sure that they don't get installed with earlier libo, then britney will hold those back [13:53] Sweet5hark, pitti, jbicha asked to get libreoffice 5.2 in for ubuntu GNOME because they use the gtk3 frontend and the current version is broken due to gtk 3.20 [13:56] hinted; but would really be prudent to fix dependencies so that -l10n cannot land prematurely [13:57] yeah, I mentioned that to Sweet5hark when l10n migrated [14:26] Sweet5hark: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#libreoffice is ignored failure now, so a RM could force/rebuild images now [14:32] pitti, thanks [14:32] pitti: thx, you rock! [14:32] I didn't do much other than bothering you with questions :) [14:34] * Sweet5hark just returned from a 30 minutes bike trip to clear my head and return some empty soda bottles. [14:36] pitti: the questions were good though: I thought about them on the bike. The hang on the autopkgtest is likely the same root cause as the hang we had during unittests (induced by CUPS somehow). I might disable cups for autopkgtests too ... [14:38] of course, the root cause in cups should be fixed triaged too -- still (re)building to get the backtrace. [14:39] pitti, seb128: who do I need to be a PITA to next for triggering the image rebuild? is there a doc/procedure for that? [14:39] Sweet5hark: ask in #ubuntu-release and ping Laney [14:39] Sweet5hark, #ubuntu-r ... what pitti said [14:40] Sweet5hark: by image, you mean Ubuntu GNOME? [14:40] or let jbicha do that since he's the one who wanted the update in for Ubuntu GNOME [14:40] well, all images that ship LibO [14:40] pitti, is https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-yakkety/yakkety/amd64/n/nplan/20160823_175352@/log.gz a known issue? should that be just retried? [14:41] seb128: and ship the beta with different versions of libreoffice and libreoffice-l10n, making it crash horribly when using a non-en-US locale? [14:41] Sweet5hark, it does crash? [14:41] but in any case Ubuntu doesn't participate in beta afaik [14:41] but yeah, +1 on getting the update in anyway [14:42] seb128: there was one report of that here in the channel by someone using a non-en-US locale on yakkety. I didnt try to reproduce -- was busy trying to fix as we had a good idea about the root cause. [14:43] the images need to be rebuilt anyway for the apt/ubiquity issue (the installer apparently fails if English isn't the default language for your timezone) [14:43] seb128: did I say already that I hate all this? :/ [14:43] * Sweet5hark <- not a happy bunny. [14:45] at least the beta will be a lot more exciting this time around :/ [14:46] Sweet5hark, oh, yeah, fun, trying to start writer on a yakkety daily lead to an error dialog indeed [14:46] that's french locale [14:47] shrug, yakkety is not in a pretty state [14:48] on edgy standards? [14:48] * seb128 fixes rhythmbox not having menus under unity [14:49] jbicha, ^ just fyi, bug #1616273 ... you seem to put a lot of work on updates, thanks for that, but maybe we should do a bit less and test a bit more before landing... [14:49] bug 1616273 in rhythmbox (Ubuntu) "Rhythmbox does not have menu" [High,Confirmed] https://launchpad.net/bugs/1616273 [14:52] seb128: likely that is just my subconsciousness making sure nobody will ever give me uploader rights. [14:53] hmm, I didn't touch menus [14:54] no, that's the issue [14:54] you updated and upstream moved to gresources [14:54] I guess you didn't test the new version under Unity though [14:58] jbicha, also the vcs and the version you uploaded have a small control diff, a line " - Context Panel" added to the upload but not the vcs, unsure which one is right [15:02] we don't ship the Context Panel so the version without that line is best [15:03] thanks, that's what the vcs had so should be what I uploaded [15:05] thanks for fixing rb! [15:05] yw! [15:12] Laney, ochosi, willcooke, bug #1616500 might be another gtk 3.20/theming issue [15:12] bug 1616500 in ubuntu-themes (Ubuntu) "u-c-c user-password-strength bars are not width enough" [Undecided,New] https://launchpad.net/bugs/1616500 [15:13] Sounds like it [15:13] They want a min-width most likely [15:14] Laney: is gtk3.20/theming likely to be the cause of the skip button wandering on the installer? [15:14] Don't know [15:14] * flocculant doesn't either - but he does know it appeared recentlyish [15:15] Laney, is that a code thing? because it works with GTK_THEME=Adwaita [15:15] Theme [15:15] oh ok [15:16] skip btn is bug 1612455 [15:16] bug 1612455 in ubiquity (Ubuntu) "Skip button misaliigned on installing screen" [Undecided,New] https://launchpad.net/bugs/1612455 [15:17] jbicha: aha - that has a dupe now :p [15:17] Anyone should feel free to try and fix either of those [15:18] well if I could - I would have ;) [15:18] same [15:18] it feels like a non trivial learning curve === sverdy is now known as StephaneVerdy [15:19] ok [15:19] the password strength might be something similar to https://git.gnome.org/browse/gtk+/commit/?id=2a0024ebe2382c94adeec7dec0ba75a8f4f4fae0 [15:20] hum, seems ours is already name levelbar as well so maybe not that trivial [15:22] we have a min-width in our theme [15:24] levelbar trough { min-width: 34px; } [15:24] back to gettext now [15:24] thanks [15:28] no problemo [15:28] I'll do some more theme stuff after getting asgen done [15:30] seb128: sorry, was at meeting; I haven't seen that failure yet, the tests haven't been too flaky so far [15:31] pitti, no worry, should I just retry? [15:31] it's one of the issues on the glib proposed candidate [15:31] seb128: (aside from regressing due to some package update which I forgot) [15:32] seb128: yes, please do [15:32] done [15:32] seb128: IIRC cjs is a regression due to gcc 5 -> 6 [15:33] and gvfs has become horribly flaky these days, not sure why [15:33] pitti, yeah, abi compliance change, jbicha reported it to debian [15:33] the cjs one [15:33] gvfs :-/ [15:33] * pitti toddles of to make dinner, bbl [15:33] we didn't change/update it this cycle though [15:33] pitti, have fun! [15:53] happyaron, nma update finally looks good, sponsored to yakkety and xenial SRU ... now what's the status of 1.2.4? ;-) [16:39] ximion: we had it good with python [16:39] Laney: what's broken? [16:39] this langpack thing man [16:39] it's the wurst [16:44] Laney: I'm not really sure how it works, but GLib should have all the i18n stuff we want [16:45] only the frontend/backend split could be pretty tricky [16:45] not really [16:45] working with random uninstalled translations is not exactly well supported [16:45] * ximion continues breaking AppStream API [16:45] I'm going to have to call localedef for every locale I think [16:46] python has a great function to just do the right thing here [16:46] oh well [16:48] Laney: does Python not compile the locales? [16:49] and even more important: they aren't they compiled already? === davmor2_ is now known as davmor2 [16:49] if we place the temporary files in /run/asgen, they could be stored in memory without the overhead of writing to disk in /tmp [16:50] seb128, tkamppeter: https://bugs.launchpad.net/ubuntu/+source/cups/+bug/1616548 <- tried to get a bit better stacktrace, but not with much success ... [16:50] Launchpad bug 1616548 in cups (Ubuntu) "Cups causes LibreOffice unittests to loop in a sbuild" [Undecided,New] [16:51] ximion: They're coming straight from the langpacks, nothing to do with the system's locales [16:51] and no, python has a .mo parser built in [16:51] maybe I could steal one of those from somewhere instead [16:54] Laney: I recently stole APTs version-compare logic, so yes, this is possible - if the .mo parsers isn't even more annoying [16:55] also: there is no C function that reads .mo files directly? [17:02] hughsie parses mo files directly to extract translation-completeness info :P [17:02] (but he obviously doesn't need to read langpack info) [17:18] night all, going to try and get the grass cut before it starts raining [17:43] Sweet5hark, you have the libcups dbg installed? weird there are no symbols [17:47] seb128: yeah, dunno. As said in the report, I installed the cups and libcups-dbgsym. I also did a "apt-get source cups" to look for a "emit.c" file, but there isnt one, just an ppd-emit.c (which has the ppdCollect2() function). that might be just usual build system (file move/rename/whatever) madness though ... [18:43] ximion: that only looks at the header [18:44] I know [18:44] meh [18:44] I'll just localedef everything [18:44] then can use gettext directly [18:54] I'll probably end up getting rid of this badass bit of code though :( [18:54] langpackLocales = path.dirEntries (SpanMode.shallow) [18:54] .filter!(f => f.isDir) [18:54] .map!(f => f.name.baseName) [18:55] .array; [18:56] food time [23:55] FWIW, I just tried a vanilla xenial->yakkety upgrade: http://pastebin.com/WjkyGtzf <- gconf2 doesnt look happy [23:59] desrt, I have a fix for https://bugzilla.gnome.org/show_bug.cgi?id=753459 - are you a good person to review? [23:59] Gnome bug 753459 in general "GDateTime: Add conversion functions from/to ISO 8601 strings" [Enhancement,New]