[06:23] Good morning [06:23] hey GunnarHj [06:28] morning pitti [06:49] good morning [08:06] bonjour didrocks, ça va ? [08:07] hey pitti [08:08] good morning desktopers [08:10] bonjour seb128 [08:10] pitti, happy friday! wie gehts? [08:11] bonjour pitti! encore malade :/ [08:11] et toi ? [08:11] seb128: gut, danke! [08:11] didrocks: eek, your knee, or something else now? [08:11] pitti: since yesterday: headache + caught a cold [08:11] so quite a slow mind… [08:14] * pitti hugs didrocks [08:14] get well soon then! [08:14] * didrocks hugs pitti back [08:14] pitti: well, better now that during the hackfest/fosdem/sprint [08:14] than* [08:15] pitti: so today: testing fsck full integration with reboot (only did running plymouthd and services manually), so with units and so on + writing man pages [08:15] so, expect a lot of reboots [08:16] (as the vms are using the ubuntu text version for plymouth, so need real hw) [08:16] didrocks: oh, really? I see the graphical one [08:16] pitti: then, everything is done apart: "non ubuntu plymouth theme and i18n" [08:16] didrocks, don't bring us the ubuflu! [08:16] pitti: squared dots? [08:16] seb128: a week should be enough to get that completely evacuated from my body! [08:17] didrocks: hm, I do see the graphical one on shutdown; booting is so fast, I actually don't see anything at all [08:17] didrocks: let me try forcing an fsck on that and reboot, that should take a bit longer, right? [08:17] seb128: oh, you're coming too? nice! [08:17] pitti: yeah should! [08:17] seb128: I hope that a week would be enough. bonus is that I'll be immunized! [08:18] pitti, well, I'm coming to Brussels for the mir/unity week after fosdem, but I'm going to be there at least on sunday [08:18] pitti: I just tried on virtualbox, I need to repair a qemu box [08:18] didrocks: and I definitively get the graphical one when booitng the live CD [08:19] seb128: ah, nice! [08:19] didrocks: well, immunized against *that particular* germ -- not the other 50 which will be floating around :) [08:19] didrocks: hm, forced fsck, still didn't see any plymouht at all [08:19] pitti: qemu is too fast [08:19] didrocks: do you start plymouth in the initramfs? i. e. do you have cryptsetup installed or so? [08:20] pitti: I don't start it during initramfs [08:20] pitti: when you say graphical -> rounded dots? [08:20] not squared? [08:20] didrocks: anyway, try various -vga options [08:20] didrocks: yes; I mean the -ubuntu theme, not the -text one [08:20] that's what you aim for, right? [08:20] ok, as most of people think about -text being non graphical :) [08:20] ah! [08:21] didrocks: apt install cryptsetup, now I see a garbled (but graphical) theme [08:21] when -text, you have the ubuntu logo, but rendered without sprites [08:21] garbled? [08:21] didrocks: well, in -text I only see a standard text cosole "Ubuntu 15.04 .." with some animated text dots [08:21] didrocks: with -vga vmware -> -text [08:22] didrocks: with -vga std -> -ubuntu (graphical) [08:22] right, the -text is http://techluminati.com/operating-systems/how-to-install-ubuntu-12-04/ [08:22] didrocks: ^ so try that one? [08:22] ah, great! [08:22] yeah, trying :) [08:22] didrocks: yeah -- any testing that involves rebooting your workstation just sucks and is way too demotivating :/ [08:23] didrocks: so: -vga std, sudo tune2fs -c 5 /dev/vda1, boot with upstart in grub -> I see graphical fsck [08:24] (percentage and stuff) [08:26] * didrocks hugs pitti, works well with -vga std :) [08:26] ok, now transferring units and mocks to the vm [08:26] and crossing fingers :) [08:27] pitti: just to keep you updated: I finished the C keystroke on wednesday [08:27] and so, it cancels all fsck processes running (and new incoming ones) [08:28] however, it's true that we have bug #776034 for now, which can be worrying on non ssd [08:28] bug 776034 in Gang Garrison 2 "failed to initialize drawing surfaces" [Undecided,Invalid] https://launchpad.net/bugs/776034 [08:29] https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=776034 rather [08:29] Debian bug 776034 in systemd "fsck runs in parallel on same physical disk" [Serious,Open] [08:29] didrocks: you (did)rock! [08:30] didrocks: btw, just making sure you know these potentially helpful tools: (1) forward VM's ssh port to the host so that you can easily ssh in, (2) enabling serial console so that you can debug stuff during boot, (3) enabling the systemd debug console [08:30] didrocks: I haven't tried combining (2) and (3), i. e. enabling the debug console on ttyS0, but that might be useful :) [08:31] I did 3, will have a look at 2 :) [08:31] well, of course, everything will work well on first boot, right? :p [08:31] didrocks: it's not much more than console=ttyS0 on the kernel cmd line [08:31] didrocks: heh yeah, what am I talking about! [08:31] hehe :) [08:32] * didrocks downloads latest vivid first and start the manpages [08:32] systemd.log_target=console console=ttyS0 [08:33] didrocks: ^ that sounds useful too, in case your units spit out some logging [08:35] pitti: thanks for the hint! /me notes that down [08:35] didrocks: http://www.freedesktop.org/wiki/Software/systemd/Debugging/ has pretty much everything, I think [08:36] pitti: yeah, that was my baseline when I needed to have a look at that. I didn't use "Debug Logging to a Serial Console" though, (until now) [08:37] didrocks: so, qemu: -serial stdio -vga vmware, and boot with console=ttyS0, then you get some crazy text plymouth on the console and then a login prompt [08:38] pitti: with qemu, you are always using the adt image, right? [08:38] didrocks: hah -- I edited debug-shell.service for s/tty9/ttyS0/, sudo systemctl enable debug-shell.service [08:39] didrocks: then just boot (no other kernel command line args) [08:39] ah sure, better that matches :) [08:39] and you immediately get a root shell on the terminal where you launched qemu [08:39] (don't use console=ttyS0 there, you can't have both the debug shell and a getty there) [08:39] didrocks: no, not always [08:40] didrocks: I do use the adt image for a lot of stuff, but for testing systemd uploads or things like that I have a standard vivid install in a VM [08:40] didrocks: I call it "pid1.img" :) it's standard vivid plus systemd-sysv installed, plus using my apt proxy [08:40] pitti: ah, I might set that up then :) [08:40] didrocks: and I always boot them with -snapshot so that my pid1.img always stays pristine [08:40] and I can hack those up as much as I want [08:40] yeah :) [08:41] didrocks: oh, and I changed grub to always show the boot menu, much nicer for that stuff [08:41] * didrocks fires an install first [08:41] i. e. comment out the two GRUB_HIDDEN_TIMEOUT lines in /etc/default/grub, then sudo update-grub [08:42] sure [08:49] morning guys [08:49] didrocks, feeling any better? [08:52] hey willcooke [08:52] morning seb128 [08:53] happy friday ;-) [08:55] \o/ [08:55] Hi pitti! [08:56] willcooke: not completely, still feeling tired and headache + sneezing [08:56] willcooke: better now than for the sprint though :p [08:56] :) [08:59] hey GunnarHj [08:59] pitti: Noticed that you changed the tag at bug #678421 to "verification-done". Did you test the gdm package for Utopic? [08:59] bug 678421 in gdm "Error message for a faulty ~/.profile script" [Medium,New] https://launchpad.net/bugs/678421 [09:02] GunnarHj: oh, I misunderstood the tag changes then, sorry; adjusting [09:02] hey hey, happy friday [09:03] pitti: Ok. gdm-trusty has been tested, but not yet gdm-utopic. [09:08] argh, kvm crashed and now, I have some hw lockup, rebooting… [09:21] morning! [09:41] pitti: it seems I can't set SOCK_NONBLOCK directly from the .socket unit, or am I missing something? (so, I would need to call fcntl() for the socket-activated path) [09:42] didrocks: I'm not aware of a way to do that; but wouldn't that defeat the purpose of activation? [09:43] didrocks: also, NONBLOCK is a client-side option, not a server-side [09:43] i. e. the activated service, nor the systemd proxy shouldn't care -- it's the client who conects to the socket who has to set that [09:44] pitti: the server can as well, to accept multiple connections without being multithreaded [09:44] (as some connections can happen after the service is activated) [09:44] take the case of multiple fsck in parallel [09:44] didrocks: ah, I'm afraid I'm not aware of that kind of usage [09:45] pitti: that's working manually (if the service created the socket), it's just that when systemd gives it to me, it doesn't have this flag, but I can set on the fd afterwards in that case, so doing that [09:47] ok, working well now with the manual non blocking flag set (when the socket is passed by systemd :)) [10:35] moin [10:36] moin Sweet5hark, wie gehts? [10:37] Sweet5hark: FYI, latest LO autopkgtest fails, https://jenkins.qa.ubuntu.com/job/vivid-adt-libreoffice/119/? [10:37] No package 'libcmis-0.4' found [10:37] it's a bit weird, as it did succeed in vivid-proposed [10:38] so this smells like a missing binary or test dep [10:38] pitti: tres bien! J'ai construit libreoffice 4.4.0.2 on amd64,i386,armhf,powerpc! [10:39] Sweet5hark: oh, on parle français maintenant ! d'accord :) [10:39] man, this channel is turning more french by the minute... [10:40] pitti: thats mostly harmless, it will go away when we bump to libreoffice 4.4. caused by a bit of hackishness in the autopkgtests ... [10:40] Sweet5hark: well, it's now holding back gtk+3.0 because of the regression [10:40] http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#gtk+3.0 [10:40] oh, and libcmis [10:40] *cough* likely means that the current LibreOffice 4.3. in main is also FTBFS though :/ [10:41] Sweet5hark: so, I think the real reason is the libcmis update to 0.5? http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#libcmis [10:41] so it's quite natural that it doesn't find -4.4 [10:41] pitti: yep, we bumped libcmis to 0.5 for libreoffice 4.4. [10:43] looks like a transition [10:43] ok, that explains at least that bit [10:43] pitti: libreoffice 4.3 should build and autotest against libmcis 0.5 too, but likely the autogenerated b-ds say it just wants 0.4 [10:46] pitti: the bad news is libreoffice 4.4 didnt yet build on arm64/ppc64el ... but on a closer inspection we never shipped those before anyway. [10:47] Sweet5hark: yeah, that shoudln't hold it back [10:49] pitti: https://launchpad.net/~canonical-libreoffice-builds/+archive/ubuntu/archive-staging/+sourcepub/4682663/+listing-archive-extra should land in the archive soonish (next days) anyway ... [10:49] cool, thanks [11:00] thats the question: whether 'tis nobler in the mind to ask for gcc-4.7 to be ported to ppc64el or to switch to ucpp making it a build-depend on all archs and thus requiring a profane and cruel MIR to make it join the jolly lot in main! [11:03] (its friday, isnt it?) === MacSlow is now known as MacSlow|lunch [12:30] pitti: ok, latest bugs squashed (a race where plymouth started after the root fsck), writing manpages now [12:30] pitti: however, in the emulator, c doesn't cancel (the signal is not received) [12:30] pitti: not sure if it's a qemu issue, I'll try rebooting my box soon [12:30] (works well with plymouth-x11 in my session) [12:31] didrocks: yay, sounds great! [12:31] didrocks: sorry, got stuck in debugging the touch regression for the past 2 hours :/ [12:32] pitti: yeah, saw that, no worry :) [12:45] didrocks: so it sounds like you might have this on the upstream list by next week, so that we can work on it on the hackfest? [12:45] oh, Netti is coming home, lunch o'clock [12:46] pitti: yeah, I'm looking at the i18n integration now, but I can post without this [12:46] pitti: basically, I think we'll send the metadata + a string [12:47] pitti: so that .script themes can use the string directly, and other themes can use the metadata [12:47] enjoy your lunch! :) === JonEdney_v2 is now known as JonEdney === alan_g is now known as alan_g|lunch [13:17] didrocks: "however, in the emulator, c doesn't cancel (the signal is not received)" phone emulator? that one does not have a physical keyboard... not allows input from it, not even if a skin with keyboard is enabled === MacSlow|lunch is now known as MacSlow [13:20] xnox: no, qemu, ubuntu-dekstop on plymouth [13:20] ubuntu-desktop* [13:20] didrocks: =( [13:21] xnox: I need to check on my system if this works though. But I guess it does (doing something similar than mountall) and it works with plymouth-x11 [13:22] and yeah, the signal is send to listen to that key, it's not something racy (as I write something on the theme screen and this part is written) [13:23] didrocks: is tty connected to plymouth or systemd. as far as i remember there are multiple things needed in the unit to have tty actually wired up to plymouth. [13:23] e.g. see oem-config / ubiquity jobs [13:23] units that is. [13:24] xnox: ohoh, nice pick! [13:24] probably plymouth tty not being connected [13:24] * didrocks looks [13:27] i see no lennart in http://lists.freedesktop.org/archives/systemd-devel/2014-November/thread.html#25288 [13:27] didrocks: ^ -enochan [13:27] xnox, look in decembre? [13:28] night [13:28] tah [13:30] xnox, night? are you in .au now?! [13:30] or is that "right"? ;-) [13:31] seb128: that was a typo for "right" [13:31] xnox: yeah, it's going on in december :) (but answered on the right chan :p) [13:31] xnox: that's why threading is bad, we need isolates! [13:42] didrocks: oh, that sounds good -- then any consumer can pick [13:43] pitti: yeah, sounds the saner approach to me [13:43] pitti: trying the TTY options for plymouth that xnox suggested, didn't get it to work for now [13:43] but looking) [13:44] didrocks: which ones? I recently fiddled with friendly-recovery and ubiquity-dm, which both want to 0wn the console [13:44] pitti: xnox told rightly that I maybe don't receive keypresses in plymouth due to StandardInput= not being set [13:44] pitti: tried to direct it to tty1, tty7… no chance yet [13:46] but for upstart we did nothing specific and the plymouthd daemon seems to fork and then set the tty itself, so, it should get the inputs… [13:48] systemd is angry and does more setup than upstart did... [13:48] xnox: well, it would as well prevent plymouthd to display the splash [13:50] didrocks: splash is output, keyboard is input. [13:51] * xnox thought there were hooks into systemd-input mechanics, unless that's for cryptsetup only [13:52] i'm so happy you are looking into this =) [13:52] xnox: yeah, but there it would be where for systemd to allow setting any ouptut and mess with inputs [13:52] xnox: ahah, I see that :p [14:00] didrocks: right, I had the same "does not receive keypresses" in ubiquity, until I set up the TTY correctly [14:02] didrocks: http://paste.ubuntu.com/9835101/ is what I did for oem-config [14:03] pitti: yeah, I tried that with tty1 and then tty7 but didn't work [14:03] didrocks: so it might be that plymouth's units need something like that (as it's plymouth which gets the keys, not fsckd, right?) [14:03] pitti: right, it's plymouth [14:03] pitti: I wonder though, because systemd as systemd-ask-password-plymouth.service === alan_g|lunch is now known as alan_g [14:04] so, it's like they already support stdin from plymouth? (this service, just run plymouth ask-question…) [14:05] didrocks: right, that's for cryptsetup passphrases and similar, so that ought to work === tedg is now known as ted [14:05] pitti: I didn't retry to reboot on my system yet, so maybe it's all working and just not in qemu for $whateverreason [14:07] headache restarting again -> fallback to manpage writing [14:12] kenvandine, hey [14:12] hey seb128 [14:12] kenvandine, seems like you registered the notes-app project on launchpad, is anyone working on it? there are some approved mps, would be nice to land those [14:13] i don't think so [14:13] i just did that as a part of mass registering all the projects [14:16] kenvandine, k, do you know who to ask about it? I'm happy to help and handle some landing/click updates (just need somebody to point me on how to do that, never did it) [14:16] kenvandine, I'm using the app and would like to maybe fix some small issues [14:17] bfiller would know [14:18] seb128, it was abandoned in favor of reminder-app [14:18] i think [14:19] kenvandine, they are differents apps though [15:09] ok, I'm going to drop from IRC, working offline in a bit and probably not joining IRC again before calling it a week, have a nice w.e everyone [15:10] cya seb128 [15:19] larsu: hi! what's the solution for apps that get black backgrounds when used with overlay scrollbars? what should I be looking for? [15:19] larsu: ie: http://snag.gy/pT2aB.jpg [15:21] mdeslaur: setting the background of affected widgets to @bg_color in the theme usually works [15:22] better is to disable overlay-scrollbars ;) [15:22] larsu: in the _theme_? so not the app itself? [15:23] ya [15:24] ah, lovely [15:25] larsu: what's the best way to turn off overlay-scrollbars in the app itself? [15:26] mdeslaur: remove overlay-scrollbars from GTK_MODULES [15:26] sorry, overlay-scrollbar (sans s) [15:27] larsu: can I do that in that app itself, or is it too late then and I have to do it in the desktop file? [15:28] mdeslaur: should be fine if you do it before gtk_init() [15:28] cool, thanks for your help larsu! [15:28] mdeslaur: yw. Sorry about the sad state of that module :/ It's a huge hack in gtk [15:29] larsu: will the new upstream stuff make it easier? [15:29] it's gone next cycle! [15:29] gone? as in ugly scrollbars, or as in the new upstream stuff? [15:29] we have ugly ones now ;) [15:30] but ya, upstream got their own version of overlay scrollbars [15:30] and we're going to be using those? [15:31] yes [15:31] cool === alexabreu is now known as alex-abreu [16:49] * didrocks goes to get some rest and sneezing in his bed. see you on Monday guys and have a good week-end! [16:49] safe flight larsu & desrt [17:51] * willcooke -> EOD [17:51] ta ta# === alan_g is now known as alan_g|EOW