[02:33] lan3y: g-s merge seems ok from my POV [06:45] good morning [06:46] Salut didrocks [06:46] salut jibel === pstolowski|afk is now known as pstolowski === ecloud is now known as ecloud_wfh [07:37] Salut didrocks, jibel [07:37] + StephaneVerdy__ [07:39] hey duflu, how are you? [07:39] good morning desktopers [07:40] Hi seb128. Going well all things considered. How are you? [07:40] I'm good, a bit tired though [07:40] I worked a bit late and the baby woke up during the night and it took me a while to manage to fall asleep again after that [07:41] just got coffee though, should be better soon :) [07:43] hey duflu, salut seb128 [07:43] lut didrocks [07:43] yeah, coffee FTW! [07:43] had an interrupted night as well? [07:44] yeah, but due to the tinitus… [07:44] and 4 nights of very light sleep in a raw isn't good [07:47] :/ [07:54] grr, spent one hour trying to reproduce a bug on cosmic iso and just realized bionic was on the usb stick :(( so stupid [07:56] well, you can write at least IT IS FOR SURE A REGRESSION :) [07:56] exactly [07:58] Which suggests you don't even notice Yaru? :) [07:59] Or lack of Yaru === lan3y is now known as Laney [08:01] moin [08:01] Laney \o [08:03] hey Laney, willcooke [08:03] how is it going today? [08:04] morning! [08:04] Too early to tell seb128 :) [08:04] :) [08:04] Sent my phone to Apple for a new battery on Monday, it's being delivered back to me today [08:04] morning willcooke, Laney [08:05] clobrano: hey, time for testing the built snap version and releasing? If you +1 on the snap, I'll upload the package as well [08:05] hey didrocks [08:05] +1 one for me [08:06] clobrano: ok, I let you releasing to the stable channel? I'm doing the cosmic side :) [08:06] didrocks: alright, doing it now [08:06] thanks! [08:07] hey duflu seb128 didrocks [08:14] Hi willcooke [08:16] hihi [08:23] good morning desktoppers [08:23] good morning all 0/ [08:29] salut oSoMoN ! [08:29] salut didrocks [08:31] didrocks: stable snap released (I'm about to write the announcement) and confirmed that the dock in bionic is now more transparent as expected [08:31] clobrano: \o/ [08:32] clobrano: yeah, confirmed as well (testing it right now) [08:32] I'll leave a note in the announcement to avoid unnecessary tickets :) [08:33] good idea ;) [08:33] (spoiler alert: you will still have some :p) [08:33] no way, not yaru is bug free :D [08:33] *now [08:33] :p [08:34] clobrano: btw, FYI: https://forum.snapcraft.io/t/snaprevs-update-failed/7106/12 [08:35] (on the snaprev regression) [08:42] Laney, saw some late-night comments on the gdm PR, thank you for digging in to last night. [08:43] willcooke, I was thinking... Ray only fixed plymouth and VT support recently. Did it regress at that point? [08:43] Maybe we couldn't tell because the bugs being fixed were at least as significant [08:45] yeah in my very limited understanding of the situation, I think there are perhaps a couple of issues, and one of those /might/ have something to do with Plymouth not giving up it's "lock"(?) on the graphics card in time [08:45] and that is perhaps coupled with a different gdm issue [08:45] but [08:45] I don't know [08:45] and I'm guessing [08:48] It doesn't call drmDropMaster in the bad case, I'm trying to figure out why [08:49] Cool. Yes SetMaster and DropMaster are nice single purpose functions. Easy to identify in theory [08:52] It = Plymouth? [08:52] sí [08:53] grazia [08:53] For the record, I haven't seen the bug in quite some time [08:55] duflu, you simply need to get a Dell Inspiron 11 and a cheap 100GB ssd [08:55] My Asus was hitting it for a week. But that was last week [08:56] That trusty Inspiron is really good at find race conditions for some reason [08:58] because it is slow maybe? I thought the bug preferred faster machines. But actually my desktop is not affected [08:58] Cosmic rays [08:59] which is even more appropriate this cycle [09:00] Cosmic rays indeed [09:00] I think it's a combination of slow(er) processor and fast disk [09:01] willcooke, although your laptop's not slow. Or was it not the X270? [09:02] That's my work work laptop [09:02] the Inspiron is my test laptop [09:24] jamesh, hey, did you see bug #1691649 ? [09:24] bug 1691649 in xdg-desktop-portal (Ubuntu Cosmic) "xdg-desktop-portal crashed with SIGSEGV" [High,Confirmed] https://launchpad.net/bugs/1691649 [09:36] duflu: the trace from the server doesn't say where and why eglMakeCurrent failed.. [09:36] tjaalton, I think I know the bug you're talking about. But are you asking me to do something? :) [09:37] I'm not affected by that one [09:37] duflu: just saying showing upstream the trace doesn't actually help :) [09:38] tjaalton, well it's not just for upstream. I like to leave myself and others a link to downstream for next time [09:43] seb128: yeah. The xdg-desktop-portal code itself looks like it should work: so there must be some reason g_dbus_interface_skeleton_export would fail without setting an error [09:43] those reports might all be from the live session [09:44] which is a special fs/env, so maybe something is not right for the portal there [09:44] there are a bunch of g_return_val_if_fail calls that could be culprits [09:46] would that show any error in the journal? [09:46] https://launchpadlibrarian.net/320147367/JournalErrors.txt doesn't seem to have anything like that [09:48] I can see some relevant journal entries in one of the more recent crashes [09:49] one of the logs has [09:49] Apr 09 12:16:37 hostname xdg-desktop-por[18168]: cannot open display: [09:49] Apr 09 12:16:37 hostname systemd[18095]: xdg-desktop-portal-gtk.service: Failed with result 'exit-code'. [09:49] Apr 09 12:16:37 hostname systemd[18095]: Failed to start Portal service (GTK+/GNOME implementation). [09:49] Apr 09 12:17:02 hostname xdg-desktop-por[18151]: Failed to create file chooser proxy: Fehler beim Aufruf von StartServiceByName für org.freedesktop.impl.portal.desktop.gtk: Zeitüberschreitung wurde erreicht [09:49] Apr 09 12:17:02 hostname xdg-desktop-por[18151]: g_dbus_interface_skeleton_set_flags: assertion 'G_IS_DBUS_INTERFACE_SKELETON (interface_)' failed [09:49] Apr 09 12:17:02 hostname xdg-desktop-por[18151]: invalid (NULL) pointer instance [09:50] "dic 20 11:27:56 hostname xdg-desktop-por[8447]: Failed to create file chooser proxy: Error calling StartServiceByName for org.freedesktop.impl.portal.desktop.gtk: Timeout was reached" [09:50] in another log [09:50] so yeah, might be the issue [09:50] unsure why there is no display though? [09:50] but I guess it could error out cleaning in those cases [09:55] * didrocks was wondering why G-S was segfaulting on dist-upgrade tests… ofc, my bionic VM was set up with 1Gb of RAM… so OOM, session closing… [09:56] gnome-boxes again? ;) [09:57] use virtualbox! [10:00] Laney, didrocks, Trevinho, do you know if gbp import-dsc is supposed to import the proper upstream commits like import-orig does? [10:03] seb128: I have no idea, but looking at the man page, it seems to confirms it doesn't even try to find a matching upstream commit [10:04] is that the gbp-import-dsc manpage you are looking at? [10:05] seb128: at least, with the low memory thing, I found some bugs leading to https://imgur.com/a/Ph724Py [10:05] that has upstream-branch/tag options [10:05] seb128: right, I think there is no autodetect, you need to explicitly set those [10:06] that's probably the upstream/latest branch and the upstream/x.y.z tag [10:06] I don't know either [10:06] k [10:06] well I tried to help Robert yesterday to "fix" the gnome-software Vcs and that was a fail [10:06] you can probably fix it up after the fact though by git merge 3.30.1 or whatever on upstream/latest, then git merge upstream/latest on ubuntu/master [10:06] to start it was missing a gbp.conf [10:06] which made gbp clone fails in fun ways [10:07] unsure why/how it worked for others who did the previous updates [10:07] but then import-dsc did import the content but the not the commits [10:07] haven't touched that one in a while [10:10] Laney, "Merge made by the 'recursive' strategy." is what it tells me after merging 3.30.2 into upstream/latest [10:10] diff is empty, status is 85 commits ahead [10:10] looks fine to me, I pushed, I hope I didn't screen up things more [10:11] did ubuntu/master as well [10:11] didrocks, Laney, thx for the help [10:11] seb128: some analysis here: https://bugs.launchpad.net/ubuntu/+source/xdg-desktop-portal/+bug/1691649/comments/7 -- if it is just live sessions the timeout might just be due to slow IO [10:11] yw, sorry to not have been more helpful though [10:11] Ubuntu bug 1691649 in xdg-desktop-portal (Ubuntu Cosmic) "xdg-desktop-portal crashed with SIGSEGV" [High,Confirmed] [10:12] didrocks, , no it's fine, I could have rfm-ed more, I just don't want to spend some hours this week to dig into documentation while we are release bugs still to work on [10:13] jamesh, can we easily increase the timeout? [10:13] jamesh, thx for the detailled comment, what you wrote makes sense [10:14] seb128: you were right about the environment looking pretty bare. That would indicate that it has been started early in the session before many env variables have been exported to the systemd user session [10:14] I'm kind of surprised there's no DBUS_SESSION_BUS_ADDRESS though [10:16] seb128: I don't see an obvious timeout argument in the called functions, but I'm not sure that would improve matters. [10:16] if xdg-desktop-portal is being started because some app wants to use it, we might just punt the timeout to the app [10:16] well, if the issue on the livecd is that the IO are too slow it might give us a chance to have it working [10:16] right [10:17] well, we can always fix the crash then [10:17] would be nicer if portals were to work on the live session but if they don't in some cases it's not the end of the world [10:40] jibel, seb128, anyone who cares: adding "modules_load=vboxvideo" to the kernel options fixes the corrupt screen in vbox issue [10:40] via x_nox and t_jaalton [10:41] tah [10:41] TIL modules_load [10:42] Laney, basically that's a cmdline way to setup modules-load.d which is processed by systemd-load-modules most of systemd-* garbage can do kernel cmdline overrides [10:42] cool [10:45] willcooke, still sounds like a local workaround and not a solution right? [10:46] they just uploaded initramfs-tools to do the central thing [10:46] that's for the installer [10:46] post-install boot still needs fixing [10:47] oic, I thought that was always working [10:47] yeah, post install seems to work fine [10:48] less the gdm issue [10:48] no it should be racy as well.. but if not then that's fine [10:49] I see the upload [10:49] nice [10:50] seb128: I've submitted a patch upstream (https://github.com/flatpak/xdg-desktop-portal/pull/236). I'll give them a chance to respond and then look at doing a fix for our package [10:50] it'd be nice to respin an iso once it's published [10:50] flatpak issue (Pull request) 236 in xdg-desktop-portal "desktop-portal: don't attempt to export NULL GDBusInterfaceSkeletons" [Open] [10:50] just in case they don't like my fix [10:51] jamesh, looks good and waiting makes sense, thanks! [11:03] willcooke: https://bugs.launchpad.net/ubuntu/+source/gdm3/+bug/1795637/comments/21 if you would like to check that please [11:03] Ubuntu bug 1795637 in plymouth (Ubuntu Cosmic) "No login screen when booting Cosmic" [High,Triaged] [11:03] Laney, I can dig it [11:03] Laney, is it in your package graveyard, or do I need to build? [11:04] ppa linked there [11:04] if you have the plymouth debug on [11:04] ?!?!? What. I clicked that link and it opened a patch [11:04] then you should get the new message "no default splash because we are already inactive" in /var/log/plymouth-debug.log [11:05] read on [11:05] something strange is going on there [11:05] the second time I clicked it I got your commetn [11:05] Maybe a double clicked [11:05] very odd [11:06] anyway, understood Laney, will test [11:06] it's in the text of the comment [11:06] thx [11:06] yeah, I must have clicked that link somehow without realising [11:06] heh [11:06] The breakfast Bloody Mary might have been to strong today [11:06] muahahhahahhaha [11:07] that's quite a step up from breakfast burrito [11:07] It sets me up for the day [11:07] which reminds me, I left some Earl Grey brewing, brb [11:08] Treat yo self [11:08] it's not tea time, it's lunch time! [11:08] * seb128 goes to grab something to eat [11:11] enjoy seb128 [11:19] seb128: I think it should load the commits yes. So it did when I was using it in the script [11:36] oSoMoN: Ping. [11:36] oSoMoN: I understand it's kind of last minute, and I apologize, but Lubuntu really wants https://code.launchpad.net/~wxl/ubuntu/+source/libreoffice/+git/libreoffice/+merge/356378 in for Cosmic. [11:37] Is that tbe right merge target? [11:37] *the [11:42] Regardless of keeping it in Git, which we can certainly handle, I can also just directly upload to the queue if you want. Up to you. [11:43] Laney, sorry to say, I can still recreate the problem [11:48] oSoMoN: Hmm, this MP doesn't seem right, but the patch is indeed what we want... [11:59] * didrocks tries to fetch more logs from gdm (using 3.30.1) [12:11] tsimonq2, looking [12:12] tsimonq2, that is indeed very late, final freeze is tomorrow [12:12] oSoMoN: Indeed. [12:17] Trevinho, hey! k, well it didn't work for some reason, sorted out by a post import merge now so it doesn't matter much === cpaelzer_ is now known as cpaelzer [12:32] willcooke: journal? [12:37] Laney, https://paste.ubuntu.com/p/2Fv7FZG94Z/ [12:37] should I turn on gdm debugging etc etc etc? [12:39] always [12:39] is it normal gdm or that patched one? [12:41] and the plmouth-debug.log would be helpful too [12:43] Laney, normal gdm [12:43] I did a fresh install and just added your ppa [12:43] Shall I add gdm too? [12:44] nein [12:44] k, turned on gdm debugging [12:45] do the plmouth one too pls [12:47] k, added to default grub and updated grub [12:47] Laney: when you say you are using nvidia, you are talking about the blob proprietary driver, correct? [12:47] yeh [12:48] ok, so running with Xorg works for you, really weird, unsure why I'm impacted [12:49] (I even didn't know that we could run wayland with nvidia, but it seems boots without this issue was running gdm gnome-shell under wayland) [12:49] yes haven't noticed any problem, sorry :< [12:51] Laney: no worry, just to double check: journactl confirms that your gdm gnome-shell is running under xorg? [12:51] Oct 08 10:32:03 raleigh.local gdm-launch-environment][4082]: GdmSessionWorker: start program: /usr/lib/gdm3/gdm-x-session "gnome-session --autostart /usr/share/gdm/greeter/autostart --debug" [12:52] ok, thanks for confirming! [12:52] * didrocks is realllllyyyyy puzzled [12:52] Laney, hopefully this is more useful: https://paste.ubuntu.com/p/WxhcNqbJ7Q/ [12:53] thx [12:53] willcooke: the plmouth log too please [12:53] oh, sorry, I thought that went in to the kernel messages [12:53] one sec [12:53] it does but you get [12:53] Oct 10 13:48:12 test-Inspiron-3137 kernel: plymouthd: 497 output lines suppressed due to ratelimiting [12:54] here we are: https://paste.ubuntu.com/p/DXpW5zYh7J/ [13:21] andyrock, you around about the g-c-c patches? [13:25] shrug, why is pulling a branch from andyrock which has one extra commit over my local/the origin one triggering a merge :/ [13:29] andyrock's branch doesn't have 6cee6fdf1022f0636d93ff673f4f2622ab2efca8 [13:34] Laney, thx, how did you find that out? (so next time I can figure it out without bothering you/the channel) [13:34] I looked at the log of both branches [13:36] I don't understand [13:36] if I do "git log ubuntu/master > ubuntu; git log azzar1/ubuntu/master > andyrock; diff -u ubuntu andyrock" [13:36] there is no - lines [13:37] https://paste.ubuntu.com/p/HPWMZTBR4n/ [13:37] FETCH_HEAD is just andy's branch [13:38] ah [13:38] my checkout was behind from one commit [13:38] I probably did reset HEAD^ one too much [13:38] Laney, thx [13:39] andyrock, do you want to rebase on the current content or should I just merge your branch instead of pulling it? [13:39] I guess both are ok way to bring the change in [13:41] yep [13:41] or you can rebase, but then when you push the merge request isn't closed automatically so you have to do that yourself [13:43] k [14:03] seb128: sorry I was out for lunch [14:03] you tell me [14:03] andyrock, how dare you eating! [14:03] :) [14:03] :D [14:03] andyrock, please pull the missing commit and rebase, I think your git foo are better than mine and it's going to be a 30s job for you :) [14:03] (let me know if it's more complicated for some reason) [14:04] also did you have other fixes coming still today? [14:05] no other fixes [14:05] okok I'll rebase asap [14:07] thx [14:09] seb128: it's just this commit https://git.launchpad.net/~ubuntu-desktop/ubuntu/+source/gnome-control-center/commit/?id=6cee6fdf1022f0636d93ff673f4f2622ab2efca8 ? [14:09] andyrock, yes [14:10] rebased [14:10] can you check now? [14:16] andyrock, better, thx! [14:16] andyrock, I'm also going to cherry pick 0c1c5570 [14:16] it's an obvious fix, even if we don't have e.u.c spike about it [14:17] seb128: yeah [14:17] it's very easy to reproduce [14:31] andyrock, uploaded [14:31] thx [14:31] I ended up with diff I didn't want in 52_region_language.patch and ubuntu-printer-support.patch for whatever gbp reason [14:31] I'm still struggling to understand the tools :/ [14:32] but they looks simple formatting changes, hopefully nothing wrong [15:06] woot, thunderbird update in the cosmic queue :) [15:07] \o/ [15:12] jbicha, hey, there has been some report (eg bug #1796451) that the mozjs/gjs update made gnome-sound-recorder, unsure if that's something you care about but I'm mentionning it in case [15:12] bug 1796451 in gjs (Ubuntu) "gjs-console crashed with SIGSEGV in g_type_check_instance_is_fundamentally_a()" [Medium,New] https://launchpad.net/bugs/1796451 [15:13] did anyone notice/see bug #1797052? [15:13] bug 1797052 in nautilus (Ubuntu) "nautilus no longer notifies when a new drive has been added" [Undecided,Confirmed] https://launchpad.net/bugs/1797052 [15:18] seb128: oh it crashes when you click Play 😭 [15:37] Laney, willcooke - by the way ubiquity looks hot in yaru [15:38] I can't claim any credit for that [15:38] tell didrocks so it can be passed on to the team :> [15:39] will surely do! madsrh is the one to congrats on this! [15:45] willcooke: there's a new plymouth in the ppa toaster, please try it when it comes out? [15:45] if you still get that "Could not find a primary drm kms device" thing, might ask if Trevinho can have a look instead [15:45] (this is a different error to the one I've always had) [15:47] xnox, 💛 -> didrocks, clobrano et al [15:48] Laney, roger roget [15:48] r [15:59] ppa ready [15:59] hm [16:00] guess I'm waiting for something to update [16:05] here we go [16:11] Laney, still broken [16:12] where should I be looking for that error, journal? [16:12] yeah [16:13] you better tell marco because it's probably not the same thing I'm trying to fix [16:13] * Laney has had 50-0 success-failure with this patch [16:13] Trevinho, need your help with this [16:13] k [16:13] Trevinho, you know the story? [16:13] the one triggered by the ssd? [16:13] yeah [16:14] dunno about that [16:14] there's probably not "the" bug [16:14] hold on, sit tight - I can't break it now [16:15] such is the nature of races, but lemme do some more tests [16:15] in this one that probably still exists mutter can't find a "primary" drm device to use [16:15] the one I was fixing with halfline was about the device being found but still claimed by plymouth [16:20] Laney, I'm about 10 for 1 so far [16:20] Keeping going [16:20] I guess there's a chance that could be the CanGraphical thing we were looking at before [16:20] you might want to add ExecStartPre=/bin/loginctl show-seat seat0 to gdm.service [16:20] ack [16:20] I'll do that now [16:20] give it another 10 tests [16:21] and then I'll turn off all the extra debugging and give it another 10 [16:21] if it is that you'd see CanGraphical=no (vs yes) on a bad boot [16:30] seb128: that rhythmbox upload, I don't understand it since gnome-shell has Provides: notification-daemon [16:35] ha, Laney I was about to declare success, and on 10 of 10 - boom [16:35] unlucky [16:36] $ /bin/loginctl show-seat seat0 [16:36] Id=seat0 [16:36] CanMultiSession=yes [16:36] CanTTY=yes [16:36] CanGraphical=yes [16:36] Sessions= [16:36] IdleHint=yes [16:36] IdleSinceHint=0 [16:36] IdleSinceHintMonotonic=0 [16:36] seb128: we need the "v" prefix for upstream-vcs-tag for rhythmbox since that's what upstream does: https://gitlab.gnome.org/GNOME/rhythmbox/tags [16:36] journal pls === pstolowski is now known as pstolowski|afk [16:36] hrm, I dont see that in the journal, I wonder if I screwed up [16:38] Laney, https://paste.ubuntu.com/p/XmMNstnJ2m/ [16:38] oh, there is is [16:38] line 942 [16:38] yeah, ok, so it's not that bug [16:38] Oct 10 17:34:33 test-Inspiron-3137 gnome-shell[738]: Failed to create backend: Could not find a primary drm kms device [16:39] OH TREVINHO [16:39] Yes? [16:39] pls help on that one if you can [16:39] ah, ok... it was just that as I imagined [16:41] apparently udev is not telling mutter about the devices or something [16:41] * Laney knows not [16:41] I was for sure getting this when also running shell from schroot when it had no access to the proper cgroup, could be that related to gdm not doing things fast enough on that side of things? [16:48] willcooke: so you're getting this quite randomly also right? [16:49] Cause logging isn't really much helping so I need to figure a way to reproduce mhmh [16:50] Trevinho, well, it worked for 9 boots, and now I can recreate it on every boot [16:51] I tried a combination of rebooting and power off and on [16:51] However, I now wonder if logging in to the session before rebooting might affect it [16:51] I was just waiting for gdm [16:51] then sshing in and rebooting [16:53] willcooke: we might see with a tmux if I can reproduce from there... [16:55] Trevinho, I can get it in to the broken state and then try a tmux [16:55] ok, suddenly it's reoccurring more frequently [16:56] Ok [16:57] brb [16:58] k, how do you want to do this? [16:58] tmate? [17:00] Yeah [17:04] tmate ppa seems broken [17:04] snaps to the rescue! [17:04] heh, snap doesnt work [17:07] trying teleconsole === Class7_ is now known as Class7 [17:22] is this a bug where you can get to login screen but login freezes everything? [17:27] bcurtiswx, no, this one doesnt even get to a log in screen [17:27] the freezing one sounds like it could be something didrocks was looking at [17:28] where cpu went to 100% [17:29] yes, sounds familiar. I'll ask him when he's on next ty willcooke [17:29] cheers bcurtiswx [17:29] Trevinho, here's a journal from when it was broken: https://paste.ubuntu.com/p/XmMNstnJ2m/ [17:29] but I dont have the x log [17:30] yeah i saw it, but there's embedded xorg log in there [17:30] I cant reproduce now [17:30] grr [17:30] It's certainly better with L_aney's patch [17:30] fixed then :) [17:30] :) [17:30] can we revert that? As having easier to reproduce helps [17:31] Trevinho, good idea [17:38] Laney, hum, k, foundations blamed rhythmbox and update-notifier as what keeps notification-daemon in main, I need to have another look later/tomorrow [17:38] just walking by the computer now though [17:38] bbl [17:51] Trevinho, ok, broken! [17:51] bah, I forgot to enable the extra logging [18:33] ok, calling it a day [18:33] see you tomorrow [19:29] Trevinho: Problem with reverting is that if there are multiple bugs you just might get the fixed one sometimes [19:29] I'd be wary of considering there to be only one problem here [19:31] jbicha, unsure why you tell me that about rhythmbox? I only did a change to the control file in that upload, no tagging/new version [19:32] seb128: I looked at the debdiff instead of the vcs [19:34] k, well you have commit access so feel free to fix :) [19:35] done [19:37] thx [20:47] Laney: do you know anything about https://code.launchpad.net/~illia-v/ubuntu/+source/ubuntu-wallpapers/+git/ubuntu-wallpapers/+merge/356434 please? [20:48] I'm not sure why the submitter added ~canonical-server. [20:48] Oh. It might be because "git ubuntu submit" defaults to that. We should probably change that. [20:49]