[06:51] good morning [07:13] good morning [07:28] Morning jibel and didrocks [07:28] good morning desktopers [07:29] Hi seb128 [07:29] hey duflu [07:30] hey hey duflu, seb128 [07:30] hey didrocks :) [07:42] hi duflu and everyone [08:06] good morning desktoppers [08:19] salut oSoMoN, ça va ? [08:19] salut seb128, ça va et toi? [08:19] reveillé un peu tôt mais ça va sinon [08:20] moi j’ai dormi comme une fleur :) [08:22] veinard :) [08:25] salut oSoMoN ;) [08:27] salut didrocks, ça va? [08:28] ça va, réveillé tôt également par contre :) [08:28] * duflu does the engineer's victory dance [08:28] Good morning oSoMoN [08:29] hey duflu, what's the dance for? [08:29] oSoMoN, I correctly predicted mutter's problem before seeing it in a profile graph [08:30] neat, congrats! [08:48] hey all [08:49] morning andyrock [08:50] Hi andyrock [09:01] morning all [09:01] hey andyrock willcooke [09:02] Morning willcooke [09:04] hey willcooke [09:04] good morning andyrock, willcooke [09:04] WHOOPS HI! [09:06] hey hey Laney [09:07] hey Laney [09:09] Hi Laney [09:10] hey didrocks seb128 duflu [09:10] how's it going????? [09:11] nice man from the water company came this morning while I was getting dressed :( [09:11] but he gave us a new shower head for free [09:12] Laney, a water saving one? Is it any good? [09:12] ya [09:12] i'll tell you tomorrow :P [09:13] he alleges that you can't tell the difference but it saves mumble litres per shower [09:13] * willcooke rubs chib [09:13] *chin [09:14] you would think the water company has interest in you using water and wouldn't give away free stuff to help you giving them less money [09:15] Laney, if unhappy you can often remove the washer from just inside the neck, making it less water saving :) [09:15] we still have the old one :P [09:17] seb128: oh, maybe that's why he told me to just leave the tap on all the time! [09:17] lol [09:17] it's more convenient to not have to turn it [09:51] jbicha, hey [09:51] good morning [09:51] jbicha, the trello board is to plan our work, I don't think it's the right place to list issues we consider important in the infra or other places [09:52] maybe I should have just mentioned LP: #1745210 in IRC instead? [09:52] Launchpad bug 1745210 in Launchpad itself "Support GNOME GitLab Issues as external bugtracker" [Undecided,New] https://launchpad.net/bugs/1745210 [09:52] I guess [09:53] unsure if you saw but it was briefly discussed on #ubuntu-devel [09:53] yes, thanks :) [09:53] np! [10:07] Laney, do you know what's the status of autopkgtests infra? is that still down due to spectrum/meltdown? or back up but with lot of backlog? [10:07] second one [10:07] queue's going down now but it's really long [10:08] k, well at least it's moving in the right direction :) [10:08] Laney, thanks [10:09] want an s390x desktop [10:09] willcooke, fyi the update-notifier/xenial SRU is getting published to udpates/security now [10:09] it's supppppppppperrrr fast [10:16] seb128, thanks. otp with jibel who said the same thing :) [10:18] k [10:49] jibel, you mentioned that the installer slides order was going to change or something around those line the other day? [10:51] seb128, yes, the keyboard page will move earlier, before the passphrase for encrypted volume IIRC [10:51] jibel, ok, thanks [10:51] seb128, https://code.launchpad.net/~xnox/ubiquity/move-keyboard-early [10:51] which reminds me to update the tests [10:53] willcooke, filed bug 1745350 about whoopsie ids [10:53] bug 1745350 in whoopsie (Ubuntu) "whoopsie id is not unique" [Undecided,New] https://launchpad.net/bugs/1745350 [12:07] jibel, if you start renaming the e.u.c registered bugs to replace ":" by "→" it would make more sense to patch e.u.c to use directly the char you want [12:16] seb128, I was thinking the same. it's the title e.u.c uses it make sense it uses the same when it creates a report on lp. Also this way line wrapping works and it's easier to read. I'll propose a patch [12:17] sounds good [14:55] willcooke: looks to me like wvdial is a dead project upstream too. [14:56] seb128: got a moment? It would seem that when I took my upgraded-from-prolly-wily laptop from zesty to artful, I lost the gui login screen... nothing on tty7 or 8, last thing is updating UTMP about runlevel change -- I'm hoping to avoid reinstalling, but need a working gui in a few hours.. [14:57] any thoughts/pointers? [14:57] didrocks: there's a few desktop universe demotions at http://people.canonical.com/~ubuntu-archive/component-mismatches.html [14:59] jbicha: any chance you've seen my issue or have $CLUE to apply? [15:00] I've heard rumors about meltdown/spectre fixes breaking stuff. Try a pre-meltdown kernel? [15:01] will do. though that would make me sad and I'd have to smack jdstrand with a trout. [15:02] smack Intel please [15:02] lol [15:03] and yeah, it's always interesting watching yet another cpu team discover the issues with opportunistic execution and how don't-care bits are not completely don't care [15:04] jbicha: so what you're saying is that do-release-ugprade to artful and then purge unity, and I should have a working system still? (modulo kernel hilarity) [15:04] cyphermox: btw, have you seen the issue where connecting to certain wifi networks gets something like "search public.intranet" appended to /etc/resolv.conf [15:04] lamont, is gdm installed/default dm? [15:04] cyphermox: and that makes things like sudo really slow until I remove that line [15:05] seb128: /etc/X11/default-display-manager contains /usr/sbin/gdm3 [15:05] and gdm3 is installed [15:05] but the dm doesn't start? [15:05] right [15:05] anything interesting in the journal? [15:05] * lamont reboots for fresh journal [15:05] jbicha, I doubt gdm not showing the greater is a meltdown issue btw [15:06] and then I'll see about kernel [15:06] * lamont will have to run away for a couple hours at around :40 [15:06] well I heard rumors about dkms troubles, I could be really wrong though :) [15:07] * lamont wonders wtf he has postgres installed on his laptop [15:08] it's tty1 with gdm btw [15:08] but if not there, yeah, check journal [15:08] jbicha: what makes no sense, unless your nsswitch.conf is messed up [15:09] in any case, never seen that, sounds like a local network config more than anything [15:09] seb128: DNS is broken during startup [15:10] cyphermox: ok, I'll try filing a bug then. I didn't touch nsswitch.conf [15:10] seb128: journal shows gdm.service has finished start-up, "... result is done" [15:10] lamont, and no gdm error in the journal? [15:11] nor xorg? [15:11] and ps shows gdm3 with a child gdm-session-worker [pam/gdm-launch-environment] [15:11] is gnome-shell installed? [15:11] no X running [15:11] 3.26.2-0ubuntu0.1 [15:11] could be normal if you are on wayland [15:12] no output for ps auxf | grep way [15:12] lamont, try maybe enabling debug in /etc/gdm3/custom.conf reboot and see if the journal log for gdm has more details [15:12] the wayland compositor is gnome-shell [15:12] so you should probably ps for gnome-session/gnome-shell [15:13] * lamont is already reboot with debug [15:14] can you share/pastebin the journal [15:16] http://paste.ubuntu.com/26458928/ <-- seb128 [15:18] you have lxc failing to start issues, dunno if that can create problems [15:18] DNS being broken is because /etc/resovlconf/resolv.conf.d/head is no longer in /run/NetworkMangler/resolv.conf [15:19] GdmSession: Setting display device: (null) [15:19] is weird [15:19] GdmDisplay: Creating greeter for (null) (null) [15:19] i915 device [15:19] how soon do you EOD? [15:20] * lamont is happy to tell GdmSession where the device is, given a howto... [15:20] and I'll be happy to take my time debugging the actual problem tomorrow... [15:22] EOD is a bit random, depends of how things go here, I might be on and off for a couple and hours and then back for a bit in the evening [15:23] I'm unsure how to debug that though :/ [15:23] thx cyphermox [15:24] the "DNS not working" is because networkmangler refuses to do what I want it to (bind developer means local named, and it's all NO I OWNS YOUR DNS) [15:24] lamont, you can try to use WaylandEnable=false in the gdm config see if that makes a difference [15:24] * lamont purges lxc to remove that factor from things [15:24] rebooting without wayland [15:25] Hi jbicha! Problem with autopkgtest of libpinyin. When testing libpinyin, for some reason it also tests libzhuyin and fails when doing so. [15:25] https://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#libpinyin [15:25] Why does it care about libzhuyin (which we should probably drop from the archive anyway)? Is there some kind of ambiguity in debian/control? [15:25] https://anonscm.debian.org/git/pkg-ime/libpinyin.git/tree/debian/control?h=debian/experimental [15:28] jbicha: I wonder about gnome-orca in universe. people upgrading having main only won't have the transition… [15:29] didrocks: ubuntu-desktop recommends 'orca' but we can explicitly seed it until after 18.04 if you want [15:30] GunnarHj: it sounds like you also need to rebuild ibus/fcitx-zhuyin then [15:30] jbicha: yeah, I'm thinking about people who removed the recommends, true that we try to add it before dist-upgrade [15:30] http://paste.ubuntu.com/26458989/ <-- seb128 [15:30] jbicha: as you wish, I don't have strong opinons [15:30] didrocks: it sounds like these people really want to break things ;) [15:30] jbicha: true enough to me, let's demote it thus [15:30] GunnarHj: but if you need an autopkgtest ignored, present your reason at #ubuntu-release [15:33] seb128: ps has no "gnome" in the output, fwiw [15:34] GunnarHj: oh that's odd, reverse-depends says ibus-zhuyin depends on libzhuyin7 but it doesn't look like it really does :| [15:34] "This traditional Chinese zhuyin input method is designed for old school users" [15:35] jbicha: haha [15:36] http://paste.ubuntu.com/26459025/ <-- now with no psql noise [15:36] jbicha: I have realized that ibus-libzhuyin and fcitx-libzhuyin should probably be built with libzhuyin13. Will try that in PPA. [15:38] Jan 25 08:35:30 rover3 gdm3[2217]: GdmSession: Setting display device: (null) [15:38] jbicha: Would that keep autopkgtest's hands from libzhuyin? [15:38] ibus-libzhuyin might not need to be rebuilt, it would be cool if somebody could test whether those 2 still work with the new version [15:38] GunnarHj: I don't think so [15:40] jbicha: So why is libzhuyin tested? [15:42] It seems to take over libzhuyin. [15:42] jbicha: seb128: pre-Spectre kernel (4.13.0-25.29) still gets Setting display device: (null) [15:43] with that, afk for a couple hours. will hope to find clue in scrollback [15:43] Laney: "Take over"? How/why? [15:43] Look at the packages it produces. [15:44] libzhuyin-dev: Development files for zhuyin input method library [15:44] probably something to do with it [15:48] seb128, thanks for accepting. I have now added some forgotten build dependencies and all is working. [15:48] I guess if the intention is to remove this package then it doesn't matter [15:49] * Laney assumes syncing from experimental was a good thing to do here ... [15:49] Laney: Right.. I'll do some more digging. (Syncing from experimental seemed to be a good idea at the time. ;) ) [15:50] tkamppeter, good! [15:55] lamont, I don't really know, you can try reinstalling lightdm instead of gdm see how that works [16:01] didrocks, are you going to do the bubblewrap MIR for bionic as well? [16:02] amano2: hum, is it filed as a MIR? I didn't know we wanted to seed it [16:02] or needed [16:03] It is filed and you were supposed to MIR it [16:03] According to the trello card [16:03] Evince and some others are patched to not use it now [16:04] But it would be an additional security layer. [16:04] amano2: ah, got trapped to not have been assigned to it: https://trello.com/c/ZMkHCrQY/20-bubblewrap-mir [16:05] amano2: assigned myself and added on my list [16:05] amano2, didrocks, that needs a security team review first [16:05] seb128: well, I'll do the normal copyright/packaging review [16:05] and then, ask security, as usual [16:06] That slipped already for artful, sadly [16:06] amano2, yeah, they are busy and we don't have an high need of that one in the default install atm [16:06] Now we have to patch the packages to be more insecure by default [16:07] well, not more insecure than it used to be, in all fairness :) [16:08] overstating issues never helps to resolve them [16:08] didrocks: did you want to put the brotli/woff2 MIRs on your list too? LP: #1742743 [16:08] Launchpad bug 1742743 in woff2 (Ubuntu) "[MIR] woff2" [Undecided,New] https://launchpad.net/bugs/1742743 [16:08] amano2, what is that impressive list of packages that is less secure? [16:08] jbicha: are you really asking for "want"? ;) [16:09] amano2, evince git has no mention of bubblewrap neither in NEWS nor configure.ac nor commit logs [16:09] didrocks: please :) [16:09] Well probably not that impressive ;) [16:10] amano2, also evince doesn't seem to be patched or to use bubblewrap, unsure where you saw that [16:10] <_amano> Nautilus, evince and maybe some others [16:11] <_amano> I can look that u [16:11] <_amano> P [16:11] _amano: my understanding is that it's just the gnome-desktop3 library that is affected [16:12] (which is stated in the MIR bug) [16:15] _amano, where is the evince code you mention? [16:16] jbicha: mind trello-carding it, add it to backlog and assign it to me, please? [16:21] thx! [16:22] <_amano> http://www.hadess.net/2017/07/security-for-security-gods-sandboxing.html?m=1 [16:23] Gdk-pixbuf seems sandboxed [16:24] Happy weekend, desktoppers ;) [16:24] Tomorrow is already Happy Friday :) [16:30] willcooke, we can dispense with gtk-mir in bionic [16:33] alan_g, you already said that to him some time ago but thanks for confirming :) [16:33] seb128, no. Two weeks ago I said it would come true in "a couple of weeks" [16:37] alan_g, ah ok, well we translated that to "we can do it", but thanks for coming back confirming we can :) [16:38] * alan_g grins: I try to ensure promises and deliveries line up. [16:43] :) [17:09] :) [18:15] seb128: just changing /etc/X11/default-display-manager to lightdm didn't kill of gdm3... [18:15] which is to say, halp how make computar werk? [18:26] seb128: hey! I built and uploaded initial normal and base bionic language-packs - I only did a quick check of the pl and en ones, but yeah, looked 'okayish' [18:27] seb128: I'll just have to check if all got uploaded that should, since last time I noticed some failed during the source-build+upload stage [18:34] seb128: since right now I need to get back to my SRU shift I'll re-check those tomorrow (since most of them didn't build yet even) [18:34] seb128: \o/ Thanks for the help. I would like to work with someone at some point (not today) on figuring out why gdm3 hates it and lightdm loves it. [18:50] good night all [18:56] lamont: I do sudo dpkg-reconfigure lightdm to switch login screens, then reboot [19:02] night all o/ [19:11] jbicha: ta. In other news, with bionic as of 2018-01-17 16:42:42 UTC, and clear back to artful, X will occasionally go into the most wonderfully strange failure mode. Mouse moves, but focus is gone into lala land, and buttons and keys are not processed in any perceptible way. [19:12] NVIDIA Corporation GM107 [GeForce GTX 750 Ti] (rev a2) with nouveau driver [19:13] note also that Wayland hates that card with a fiery passion (maybe), gdm3 certainly does. So that machine is using lightdm since the change to gdm3 in artful. I suppose I could check on that in the bionic world... just not today. [19:16] lamont: you don't have both lightdm and gdm installed at the same time do you? I've found issues with that in the past [19:18] gQuigs: I do, but it's easy to switch, at least until gdm3 upgrades and switches it and I have to undo it (maybe) -- tever [19:18] lamont: purge one completely and see if things improve [19:18] the X lockup is one of those frustrating "all I've figured out to fix it is 'hit the reset button'" kind of things [19:19] gQuigs: I reinstalled gdm3 only because ubuntu-desktop depends on it. not interested in removing it, and it's clearly orthogonal to X not liking life at random times. [19:20] lamont: the 'equivs' package can apparently fake up deps [19:24] lamont: I wouldn't be so positive about the clearly orthogonal part, try purging lightdm then.. [19:27] gQuigs: well, we know that gdm gets completely lost. [19:28] but yeah, today's pain is sorted "it works". Tomorrow or later is the "let's skip productivity in the name of debuging bionic" [19:28] because of primary desktop [19:30] I've found previously that their ability to get lost is much higher when there are 2 dm installed, even if one is supposedly disabled [19:36] jbicha: I failed to build ibus-libzhuyin and fcitx-zhuyin with the higher version of libzhuyin-dev (there seems to be more into it). So, to not break those input methods (for the case they work properly; I'm not sure of that) one idea is to exclude the libzhuyin-dev and libzhuyin13 binaries when building libpinyin. I'm about to test that in the PPA. [19:36] https://launchpad.net/~gunnarhj/+archive/ubuntu/ibus-libpinyin/+packages [19:36] Does that make sense to you? [20:04] GunnarHj: yes I think we could do that temporarily. Could you ask Debian why it's still in experimental and if they know about those rebuild problems too? [20:07] jbicha: Yes, I'll ask someone about that.