[07:21] msg MemoServ READ NEW [07:21] sorry guys [07:21] retro [07:25] :) [07:32] wow! [07:32] morning :) [07:33] laaaaaaaaaaaarrrrrrrrrssssssssssss [07:33] Laaaaaaaaaaneeeeeeey [07:33] oh, hey larsu [07:33] how are you? Enjoying the sprint? [07:33] we were just speaking about you [07:33] hi seb128! [07:33] oh really [07:33] and ice cream! [07:33] * larsu didn't do anything [07:33] how is Berlin? [07:33] oh. ice cream. where? [07:33] right, which is the issue :p [07:34] seb128: not warm [07:34] rainy [07:34] we are looking at somebody wanting to update gtk and fix the themes :p [07:34] * seb128 hides [07:34] joke aside we miss you here :-( [07:34] * larsu laughs and laughs and laughs and laughs and starts crying a little bit [07:34] seb128 will pay himself [07:34] €1000 per changed line [07:34] that better is a short diff [07:34] not sure it's worth it :) [07:34] wait [07:34] WAIT [07:35] haha [07:35] I've an idea [07:35] attente, if you update gtk I might get you out of the Vancouver sprint [07:35] D: [07:35] there's a Vancouver sprint? [07:35] snappy team one [07:36] so tempting [07:36] that's just around the corner for attente. Same country! [07:36] hi attente! [07:36] yeah, he got invented [07:36] hi larsu! [07:36] he's thrilled [07:36] invited even [07:36] I can imagine [07:36] vancouver is a nice city I hear [07:36] i thought willcooke was our themes guy [07:36] somebody told me that it got more rain that London [07:36] :D [07:36] google seems to back that statement up [07:37] oh, right! [07:37] seb128: is that possible? [07:37] that means it must be under water the whole time [07:37] apparently [07:37] it rains like 150 days a year [07:37] * larsu does some calculations and comes to the conclusion that that's a lot of days [07:41] larsu: how's cockpit? [07:41] Laney: loads of fun [07:41] tons of new stuff still [07:41] * larsu is having a blast [07:42] we publish to a debian repo and ubuntu ppa for every release now [07:42] \o/ [07:42] and people seem to already be starting to use it [07:42] nice [07:42] ya, my first little project :) [07:43] Laney: how's gnome software? [07:43] did you guys settle on a name? [07:44] * larsu should download the release [07:44] getting there [07:44] it's a little bit rough in the release [07:44] but we fixed some stuff since and there is an sru waiting [07:45] nice [07:45] attente is rebasing on 3.20.2 now [07:45] congrats! [07:45] the name is Ubuntu Software [07:45] * Laney coughs [07:45] makes sense [07:45] Ubuntu Software Center was perfect [07:46] but changing the name made sense, since it's all new [07:50] http://people.canonical.com/~laney/weird-things/lifeisgood.png [07:50] awww [07:53] cool eh [08:00] indeed [08:21] desktoppers: please create yourself a board under this group: https://trello.com/ubuntudesktop [09:09] wakkety will! [09:12] xnox: hey... You remember https://code.launchpad.net/~xnox/upstart/lp1433013/+merge/259317 ? [09:13] xnox: since I've been just facing with a bug with gnome terminal missing the menu-bar... This is because it dbus-activates the gnome-terminal-server, and that instance was missing the proper GTK_MODULE env var [09:13] Trevinho, yeap, same bug [09:13] but that part of code never worked though =( [09:13] despite all that code, it fails to actually do anything =) [09:13] so... Ok, that bug can be even fixed in a different way, but... It's just an example how we should make upstart and dbus match the env [09:13] xnox: well, patching upstart with it, it does work [09:14] Trevinho, oh really, i must have done it wrong. [09:14] I've just taken the xenial src, patched with it... and voila! [09:14] Trevinho, maybe i should really land it, and SRU it? [09:14] xnox: yeah, it would be nice [09:14] Trevinho, do we need it on touch too, or not? [09:15] xnox: no, not that I know [09:15] ack [09:15] i still see some alt+f2 / super differences with that, but still better than nothgin [09:15] nothing* [09:19] xnox: it could be probably nice to do the other way around, so... Dbus vars to be set to upstart [09:19] willcooke: I'm struggling to remember... Why do we use Trello? Other than for cosmetic superiority over Launchpad? [09:21] Trevinho, yes [09:21] Trevinho, but patching dbus is not nice either =( [09:21] Yeah, I guess [09:22] xnox: then i had this for gnome-session, which could be sitll be worth http://pastebin.ubuntu.com/16092825/ [09:22] if only there were hooks =) [09:23] Laney: FYI, I committed https://git.launchpad.net/~ubuntu-release/+git/autopkgtest-cloud/commit/?id=eda12ffa to get some less spam; but it'll be a while until I can roll this out as upstart doesn't use the new job file as long as there's still any instance running [09:23] Trevinho, working out which environmental variables are missing from where, and what should be the environment variable flow would be a good thing. [09:24] pitti: nice --- do you have a way to monitor for stuck instances? [09:30] Laney: what do you mean with "stuck"? adt-run has timeouts for pretty much everything, so if an ssh command doesn't come back after a reasonable time it is aborted and counts as tmpfail [09:30] duflu, hey, do you think you can manage to have a look to that upower/usd issue? [09:30] pitti: quota exceeded is due to nova messing up usually right? [09:31] seb128: Only on weekends. It's my production server :( [09:31] Annoyingly its sister machine always runs the latest Ubuntu and has no problem [09:32] k, sorry to ask you to debug that, but nobody here seems able to reproduce [09:32] Laney: in a lot of cases, yes; over time, the quota DB gets out of sync, and it says that I use e. g. 5 instances, 4 CPUs, and 25 GB mem more than I actually do [09:32] Laney: so from time to time I need to prod IS to reset that [09:32] pitti: so if you're filtering out those mails then it might be that nobody knows about the quota errors for a while [09:32] seb128: Well, there are 30-something people subscribed to the bug (from the 4000 crash reports) [09:32] pitti, does the udisks issue duflu emailed us about ring any bell to you? [09:33] Laney: but also, if one more more "bigtests" run (which use m1.large instead of the autopkgtest flavor) these block more resources, and hten we can't run the full number of workers any more [09:33] duflu, right, I know, I would like to see it resolved as well [09:33] seb128: upower? no, not off the top of my head [09:33] seb128: I'm guessing it will be "query this DMI thing", which I can do right now in 14.04 [09:34] pitti, right, wrong u, sorry ;-) [09:35] Laney: and while I'm at it, "No valid host was found. There are not enough hosts available." should also be blacklisted I guess [09:36] seb128: I'm guessing my desktop has no "lid" so the object is NULL for a reason [09:36] duflu, well, it shouldn't have a lid if it's a desktop right? [09:36] seb128: Not sure. There is a chassis switch [09:37] which may qualify [09:39] that might be what is specific to your config [09:42] seb128: I'm more curious about why the absence of a lid switch is a fatal error [09:43] That's what it sounds like [09:43] bug I guess [09:45] pitti, do you know offhand if there is a way to tell libc to read locales definition in another path than /usr/lib/locale/locale-archive ? [09:46] seb128, duflu: https://git.gnome.org/browse/gnome-settings-daemon/tree/plugins/power/gsd-power-manager.c#n2589 [09:46] this does an explicit check for the lid [09:46] which u-s-d doesn't according to https://launchpadlibrarian.net/241244694/StacktraceSource.txt [09:46] * pitti follows up to the bug, better place there [09:46] pitti, danke [09:47] * duflu grabs the xenial source to compare [09:49] pitti, seb128: Indeed the xenial source is missing the manager->priv->lid_is_present check [09:50] followed up to bug 1546641 [09:50] bug 1546641 in unity-settings-daemon (Ubuntu Yakkety) "unity-settings-daemon crashed with SIGSEGV in up_exported_daemon_get_lid_is_closed()" [Critical,Confirmed] https://launchpad.net/bugs/1546641 === Drac0 is now known as Guest53461 [09:50] seb128, duflu: more importantly, lid_is_present seems to be missing completely from u-s-d? [09:51] pitti, duflu, that was added in https://git.gnome.org/browse/gnome-settings-daemon/commit/?id=d491ebf6325fd1883e5fa81ee0f5daa7eb281c3b [09:51] pitty: Horray [09:51] then we might need https://git.gnome.org/browse/gnome-settings-daemon/commit/?id=dcf8dc956b050dc0e246d48ee7365d88a8f09357 [09:52] wow, u-s-d hasn't been merged with g-s-d in a long time, has it [09:53] indeed not [09:53] well some of the plugins were [09:53] but not the power one [09:53] andyrock: https://code.launchpad.net/~3v1n0/bamf/safer-dbus-runner/+merge/293229 [09:53] I started having a look some cycles back but it was quite some changes/not easy to review and we don't really have issues justifying it [09:54] well out of that one now [09:56] seb128: where did you push that snapcraft yaml yesterday? I'd like to take a look [09:56] pitti: stress-ng seems to be stressful :P [09:57] popey, nowhere yet [09:58] seb128: oh, i thought you put it somewhere for sergio to look at? [09:58] pitti: The NULL is not coming from u-s-d. It's originating from the upower logic [09:58] 'proxy' is NULL [09:59] So the absence of a lid may be fine. Just some UPower bug [10:00] seb128: anyway, can I get a copy? :) [10:00] duflu: proxy? the only NULL thing in https://launchpadlibrarian.net/241244694/StacktraceSource.txt is the internal object= in the topmost frame [10:00] as there is no lid [10:00] popey, you need it now, I'm cleaning and was aiming at pushing before lunch [10:01] up_client_get_lid_is_closed() only returns a bool, there is no third value to say "there is no lid", you need to check it before [10:01] seb128: ok, I can wait :) [10:01] pitti: Yeah I'm looking at the xenial source. It's not up_client that's NULL... it got past that [10:01] duflu: exactly [10:01] Admittedly I have no idea what proxy is [10:01] And now it's dark and dinner time [10:03] pitti: The NULL object appears unrelated to lids of any sort. It's just a context pointer [10:03] Laney: yeah, exit code 255, apparently it kills ssh :/ [10:03] UPower client not initialized properly [10:04] I thought you just said that up_client was not NULL (and that's also what the stack trace says ) [10:04] it does call the up_client_get_lid_is_closed() internal function, thus both the proxy and the up_client are fine [10:05] pitti: No, it crashes on the first attempt to dereference proxy [10:06] ok, then I don't know [10:06] It's in UPower I think [10:06] And not related to lids [10:06] and it seems you have some different bug than the one above [10:06] I'm just going on the stack trace in the bug [10:06] popey, try http://paste.ubuntu.com/16093118/ (sorry, having > 3 conversations ongoing) [10:07] \o/ [10:07] thanks [10:07] popey, oh, change the +exec /bin/bash #$SNAP/usr/bin/gnome-calculator back :p [10:07] popey, I made it start bash for debug [10:08] what is that a diff from? [10:10] popey, sorry, lp:~ubuntu-desktop/+junk/gnome-calculator-snap [10:11] thanks [10:13] seb128, pitti: Any chance this is just GSD starting before UPower? [10:13] That's what it feels like [10:13] A race [10:13] could be [10:13] USD starting before UPower [10:13] what is weird is that usd didn't change for cycles [10:13] why would that start being an issue now? [10:13] Yeah but raciness can change with the rest of systemd [10:14] Should be easy to test. If only I still had a xenial install on this machine [10:15] so you basically say to try to start u-s-d without having upower [10:15] ? [10:15] seb128: Yep, try it [10:15] duflu, seb128: upower is dbus activated, that can't happen [10:15] connecting to o.f.UPower will block until it's running [10:16] Hmm. [10:16] pitti, seb128: OK too hard. EOD instead... [10:16] o/ [10:16] duflu: I'd really recommend either adding that commit or replacing all _lid_is_closed() calls with "False" for a local test if it stops crashing [10:16] I'll need to reinstall xenial out of hours... [10:16] pitti, re. the previous comment, up_client_get_lid_is_closed() shouldn't segfault on a system without lid I would argue [10:17] duflu, have a good evening! [10:17] Later [10:17] seb128: yeah, an assertion failure would be more adequate [10:17] from a programmer's POV [10:17] or just returning FALSE [10:17] not that it would change much for the user [10:17] seb128: no, that'd be wrong [10:17] you'd need a third state [10:17] at most it could return True [10:18] right [10:18] you mustn't suspend a desktop without a lid, thinking that the lid is closed :) [10:18] Laney: https://code.launchpad.net/~3v1n0/gnome-session/upstart-export-variables/+merge/293236 [10:18] so it could be changed to return an enum, but that'd be an API/ABI change and even more intrusive [10:18] let's see if duflu manage to debug a bit more tomorrow [10:18] making it an assert would make the error a bit clearer, though [10:18] if that's the issue [10:18] I wonder what old upower did/if that changed [10:19] because at said the client side code didn't change [10:19] I guess I can try on a porter box [10:19] those don't have lid [10:20] pitti, thanks for the input on that bug [10:20] pitti, unsure if you saw my other question but do you know if " /usr/lib/locale/locale-archive" is easily relocatable? [10:20] (I guess not) [10:21] seb128: hmm, $LOCPATH I think, but I'm not sure if that applies to the archive as well [10:22] pitti, let me try that [10:22] seb128: see man locale [10:22] seb128: but I only ever tried that with plain compiled locales, not the archive [10:22] that would work too [10:22] seb128: not sure what you want to do, but you don't necessarily need an archive format? [10:23] I want to do an hack you don't want to know about :p [10:23] basically trying to get a snap to load translations, and since there are no locale in the core trying to include those in the snap... [10:23] * seb128 ashamed [10:34] * dpm hugs seb128 [10:34] * seb128 hugs dpm back [10:35] seb128: I thought snaps did all those magic bind mounts [10:36] we still can't build snaps with normal install locations? [10:36] dunno about the mounts but if they do that doesn't include locales [10:36] and no [10:36] seb128: no, you were right, I don't want to know :) [10:36] lol [10:36] * seb128 hugs pitti [10:37] * pitti mumbles something about "btrfs subvolumes", a solved problem for many years, and goes on :) [10:38] one day btrfs is going to be good enough to be default :-) [10:38] seb128: for this purpose it's been good enough for years [10:38] this is all read-only after all [10:39] righ [10:39] right [10:39] much bettern than overlayfs for sure [11:12] skip btrfs. straight to zfs. [11:13] * popey hugs btrfs tightly [11:14] has been chugging along happily on my laptop since I have it (~ 2 years), *shrug*; but I do realize that cking is able to break it [11:14] then again, show me a thing that cking cannot break :) [11:14] * cking likes testing stuff [11:14] cking: for example, stress-ng kills ppc64el instances in a loop (more precisely, it seems to kill sshd) [11:16] pitti, stress-ng is quite good at finding issues like that, but killing sshd is a bit wrong [11:16] cking: http://paste.ubuntu.com/16093758/ is the logtail [11:16] cking: 255 is ssh's exit code for interrupted connections usually [11:18] willcooke: https://bugs.launchpad.net/ubuntu/+source/glom/+bug/1573801 [11:18] Ubuntu bug 1573801 in glom (Ubuntu) "glom: error while loading shared libraries: libboost_python-py34.so.1.58.0: cannot open shared object file: No such file or directory" [Undecided,Confirmed] [11:18] looks like it needs a rebuild? [11:19] pitti, what does the kernel log say, was the kernel killing the entire session because stress-ng was too stressy for it? [11:20] cking: I don't have it right now, I'll see to run this manually and grab it and file a proper bug [11:20] cking: doesn't seem to affect x86 [11:20] but, lunch first :) [11:20] popey, looks like it, can you tell doko? ;-) [11:20] pitti, indeed food is always best [11:21] popey: bug 1573801 [11:21] bug 1573801 in glom (Ubuntu) "glom: error while loading shared libraries: libboost_python-py34.so.1.58.0: cannot open shared object file: No such file or directory" [Undecided,Confirmed] https://launchpad.net/bugs/1573801 [11:21] Murray also G+'ed about it [11:21] glom is an orphan, nobody maintains it in Ubuntnu [11:23] he should snap it ;-) [11:25] el troll [11:27] * xnox ponders what is glom [11:33] willcooke: un petit bug pour vous https://bugs.launchpad.net/ubuntu/+source/software-properties/+bug/1576167 [11:33] Ubuntu bug 1576167 in software-properties (Ubuntu) "Percentage bars are ugly" [Undecided,New] [11:52] oh, popey filed another dup bug ;-) [11:52] * seb128 closes and collect free lp karma [11:52] :( [11:53] popey, sorry man :-) [11:53] willcooke can still fix it! [11:53] \o/ [11:53] fix it wiiiiilllll [11:55] will'll fix it for you, and you, and you [11:57] davmor2, let's have a look in Willy's magic chair... [12:03] pitti, the LOCPATH tips works, is there a similar one for the .mo to load? TEXTDOMAINDIR seems to work for the gettext command but only for that === alan_g is now known as alan_g|lunch [12:06] seb128: hmm, bindtextdomain() doesn't say, but I thought there was some env variable, han gon [12:06] pitti, NLSPATH? [12:07] seb128: sounds plausible, yes [12:07] pitti, thanks [12:08] seb128: indeed, in man catopen [12:12] popey, seb128 - looks like a very similar bug to the calendar white corners. If I remove the view class it works. [12:18] willcooke, https://bugs.launchpad.net/ubuntu/+source/ubuntu-themes/+bug/1052936/comments/5 might be useful? [12:18] Ubuntu bug 1052936 in ubuntu-themes (Ubuntu) "Progress bar in "Progress" section has a hole in it" [Medium,Triaged] [12:18] unsure if it's a theme or a widget bug, maybe we should set some bg [12:19] seb128, thanks. yes, I think that's exactly it [12:49] desrt: https://bugzilla.gnome.org/show_bug.cgi?id=765723 [12:49] Gnome bug 765723 in gio "g_desktop_app_info_new_from_filename returns NULL if the executable does not yet exist." [Normal,New] [13:01] no shit [13:01] andyrock is andrea? [13:01] how did i not already know this? [13:01] lol [13:01] everything makes so much more sense now === alan_g|lunch is now known as alan_g [13:20] pitti, NLSPATH doesn't seem to work, seeing that we patch the glibc code to add the langpack dir I'm going to assume that there is no dynamic way (though we might have done it this way to get to work in any case and not relying on an env to be set?) [13:20] anyway I don't need to get that working today [13:20] I was just playing with it to see if I get somewhere [13:22] grr.. run into a nasty bug with unity [13:24] unity 7.4 - unable to interact with windows (gnome terminal 3.18) after a while [13:24] you have a more general problem there: apps give hard-coded paths to bindtextdomain() as a matter of course [13:24] so even if you managed to get the system message catalogs working properly, pretty much every app using gettext is gonna have a bad time [13:25] going to need to "fix" bindtextdomain() :p [13:25] or snappy... [13:25] you would think so [13:26] we should be using mount namespaces [13:26] this /snap/ business is a mountain of pain [13:28] another alternative is to recompile things [13:28] the core issue here is that we're using .debs that assume that they are installed at a particular point -- and they are not [13:30] just use jhbuild for gtk snaps ;) [13:30] won't help you with glibc :p [13:30] but ya... otherwise, it's actually a pretty nice idea [13:30] put glibc in jhbuild!!!!!!!111! [13:30] hah [13:31] this is why larsu doesn't work here anymore [13:31] https://answers.launchpad.net/unity/+question/292688 [13:31] anybody else having same issues? [13:31] a1fa, hey, I don't think that was mentioned/reported before [13:32] thanks [13:32] hikiko, andyrock, Trevinho: ^^ [13:32] it's annoying as hell, seb128 anything i can collect from unity to post/provide? [13:32] a1fa, the people desrt just pinged might be able to help you [13:33] official response from unity developers: "let's remove gnome-terminal" [13:33] :) [13:33] lets do it [13:33] but what do we replace it with? [13:33] terminator, apparently [13:33] vt1 [13:33] the terminal-app snap indeed [13:34] hm [13:34] term.js!! [13:34] a terminal is just about the last thing that one wants to snap [13:35] so desrt i sthat really the official answer? bundle something else? [13:35] a1fa: na.. we're joking around here [13:35] i am all for it [13:35] we're all at a sprint together right now [13:35] hey a1fa [13:35] yello' hikiko [13:35] they're speaking italian now.. the only thing i understand is the word "terminal" from time to time [13:35] a1fa: can you please open a bug [13:36] he already did andyrock [13:36] a1fa: i actually got once... [13:36] https://answers.launchpad.net/unity/+question/292688 [13:36] hikiko: that aws a support request... [13:36] not an answer [13:36] a bug report it's easy to track [13:36] sorry :) [13:36] https://bugs.launchpad.net/unity/+bug/1576224 [13:36] Ubuntu bug 1576224 in Unity "Unity 7.4 - bug - unable to interact with windows after some time" [Undecided,New] [13:36] also try with sudo apt-get remove gnome-terminal should fix the issue [13:36] ;) [13:37] this is new, i used 15.10 before unity 7.3? and it never happened in 7.3 [13:37] I don't use gnome terminal but I am going to try to reproduce that issue [13:37] thanks for reporting a1fa [13:37] just kidding we are on it [13:37] you guys are great [13:37] you fixed my last two bugs with unity and gnome terminal [13:37] a1fa: when you reproduce it try to switch tty [13:38] i don't use gnome-terminal too much but i noticed something wrong with the mouse and gnome-terminal [13:38] usyally a tty switch works here [13:38] alt+f1? [13:39] a1fa, ctrl+alt+f1 and then ctrl+alt+f7 to go back [13:39] yeah did that [13:39] did not fix [13:39] only way to fix it, is to "exit" [13:39] I had some issues with the ttys lately :/ [13:40] now, when i switched back, my unity bar was all garbled [13:40] but thats a whole another story [13:40] no it's not [13:40] and i moved my unity bar to the bottom :) [13:40] what do you mean garbled? [13:40] oh [13:40] a1fa, [13:40] typical gpu artifacting [13:40] did you dist-upgrade when you got that issue? [13:41] i had to do dist-upgrade to go to 16.04 [13:41] the question is: [13:42] it was after a dist-upgrade (after which you didn't reboot) that you got your unity bar garbled? [13:42] oh no [13:42] ok :) [13:42] if i take a screenshot of the garble, it disappears [13:42] oh [13:42] but also hovering over the menu, clears it [13:43] then it's the redraw-damage issue I am trying to fix for the ezoom ... :/ [13:43] I'm investigating that a1fa [13:43] i'm on nvidia gtx970 [13:43] if that would make any difference [13:43] no :) but thank you [13:44] cool :) [13:44] opening the menu causes a redraw [13:44] and solves the problem [13:45] but I'm trying to do a proper fix for that [13:45] so that you don't have to :p [13:46] how do you guys use your unity launcher? bottom? left? auto hide? [13:46] i have a hard time adjusting to auto hide, can never get sensitivity right [13:46] I use it on the left and always visible [13:47] oh, i had another thing.. mouse acceleration [13:47] I like it to be visible all the time because the wide screens have all that space on the right [13:47] turning down mouse pointer speed does not disable mouse acceleration tottally [13:48]