[01:21] Laney: yeah, I an see that === pstolowski|afk is now known as pstolowski [07:16] good morning [07:17] salut didrocks [07:17] salut jibel [07:18] didrocks, en forme ce matin? [07:21] jibel: ça va, toujours à chasser mon bug gdm par contre [07:21] et toi ? [07:24] Good afternoon didrocks, jibel [07:25] hey duflu :) [07:33] Hi duflu [07:33] didrocks, ça va bien, toujours à chasser les bugs uefi et de shutdown [07:34] pas les plus funs à troubleshooter [07:35] non et c'est uniquement sur du hw donc pas de snapshot et il faut tout reinstaller à chaque fois [07:39] ouais, voilà… [07:39] pareil pour moi et gdm, je suis obligé d'aller dans des sessions "cassées" pour debugger [07:39] (et uniquement sur ma machine apparemment :/) [07:44] good morning desktopers [07:44] lut didrocks jibel [07:44] good morning desktoppers [07:45] lut oSoMoN [07:45] en forme ? [07:46] Hi seb128, oSoMoN. How goes? [07:47] salut seb128 [07:47] hey oSoMoN [07:47] I'm good, thanks! [07:47] how are you all? [07:47] good, full night sleep without interruption ftw! [07:47] sorry, shouldn't say that in front of you [07:47] those are the best :) [07:59] * willcooke wonders if T_revinho is actually a vampire [07:59] or at the very least nocturnal [08:00] night is the best time for quiet hacking right? [08:00] hey willcooke [08:00] :) morning seb128 [08:00] did he send patches at 3am again? [08:00] hey willcooke [08:00] 03:51 [08:00] hey didrocks [08:01] Morning willcooke. Yeah it was mid-morning here and he was still on IRC, again [08:01] heh [08:02] hey willcooke [08:02] yo [08:02] Trevinho, sleep well man! [08:02] yo Laney [08:02] hey Laney, how are you? [08:02] hey Laney [08:04] hey oSoMoN seb128 didrocks [08:04] i'm alright, been pruning in the garden this morning 🤘 [08:04] what's up [08:04] nice [08:05] Umm, wutup Laney? [08:05] certainly not the volume of the clematis [08:05] MUHAHAHAHAHAHAH [08:06] \o/ [08:06] those things would invade the house if they could [08:07] I was going to prune the Buddleia the other day, but the internet told me not to until spring [08:07] was a close call, I had the loppers ready to go [08:07] in the spring I discovered that one of them had gone over the back wall and half way along the neighbour's washing line [08:07] true story [08:07] hah [08:08] * duflu does love the electric pruner. And no I won't do yours... [08:08] we don't have space for one of those [08:08] my next acquisition is going to be a few cannas [08:09] for containers [08:10] Gardeners' World had cannas on the other day [08:10] yep [08:13] only thing I'm worried about is them being nicked [08:13] as I'll have them out the front probably /o\ [08:13] that's the south side [08:14] hm, tricky [08:15] Dont really want great big chains round the pot, because then they will certainly be stolen [08:15] also ugly [08:21] willcooke, are you moderator/Admin on the hub? can you delete the spam posts from https://community.ubuntu.com/u/4bd2/summary ? [08:21] erm [08:21] so, [08:22] yes it seems I can [08:22] and I just deleted them [08:22] thx [08:22] but I didnt mean to [08:22] oh well [08:22] unsure if we should ban/block that user [08:22] problem solved [08:22] it was just spam anyway [08:22] right [08:22] can you see if the spam posts are gone too? [08:22] I wonder if I need to delete those as well [08:22] they are [08:22] thx [08:22] woo [08:23] * willcooke <- super cow powers [08:23] :) [08:25] Laney, btw did you end up accepting rhythmbox or was it someone else? [08:25] seb128: not me [08:25] :( [08:25] to be fair if I wanted to reject it I should have [08:25] sorry, I didn't have time to address your comment before it went accepted [08:26] I'm not sure how component-mismatch deals with provides [08:26] did it fix the problem? [08:26] maybe we could have demoted notification-daemon without change to the recommends? [08:26] I'd have thought so [08:27] recommending a notification implementation was probably correct [08:27] it's not a very big deal [08:27] right [08:27] well suggest/recommends, depends how much you think notifications are part of the experience [08:27] but yeah [08:27] it didn't solve it because update-notifier has a similar Recommends [08:27] let me try to demote n-d and see if component-mismatch complains [08:28] I'm unsure if there is a "dry" way to try that? [08:28] where is the problem visible? [08:28] https://people.canonical.com/~ubuntu-archive/component-mismatches.txt not there [08:28] nowhere [08:28] notification-daemon is in main [08:28] someone randomly noticed? [08:28] but we believe it should be in universe [08:28] it was part of the "has no owner" packages list [08:28] oh ok [08:28] foundations worked making sure all packages in main have an owner [08:28] well kick it out imho [08:29] right [08:29] and see if it complains, I don't expect it to but we'll know for sure then [08:29] just foundations said that rb/u-n needed fixed first [08:29] but maybe they are wrong/didn't notice the g-s provide [08:29] right [08:29] doing that now [08:29] it's easy to overlook those [08:30] https://people.canonical.com/~ubuntu-archive/germinate-output/ubuntu.cosmic/rdepends/notification-daemon/ [08:31] demoted, let's see [08:31] willcooke: you could probably check vbox today btw, the iso should have the fixed initramfs [08:31] Laney, sweet [08:31] downloading [08:31] I need to test Hyper V as well [08:31] * seb128 does that as well [08:32] Windows 10 has ssh and scp built in [08:32] neat [08:33] Oh, another true story: [08:34] I sent my phone away for a new battery. Before it went I had to install itunes on the windows box to back it up. So far, so good. Phone came back yesterday and I had to plug it in to iTunes again to restore the backup. Only iTunes now won't run. The machine has done *nothing* since I backed the phone up on Sunday., and suddenly it just doesnt work [08:34] rls-w10-incoming [08:34] :) [08:34] I had to uninstall and reinstall itunes [08:34] then it worked [08:34] ?! [08:35] Computers are a silly thing [08:35] That's the end of Story Time for today [08:41] seb128: c-m is probably right saying that mozjs52 needs demoting btw, not sure if you wanted to do that? [08:42] why is that in c-m-proposed only? [08:42] oh, it's not [08:42] http://people.canonical.com/~ubuntu-archive/component-mismatches.svg is misleading [08:42] think the svg only shows the other way [08:43] right, looks like it [08:43] I deleted mozjs38 yesterday btw :) [08:43] back to having only 2 mozjs in the archive [08:43] Laney, demoting done [08:43] yeahhh [08:44] willcooke: closed source, not computers in general ;-) [08:45] :) [08:46] Laney, jibel tjaalton - Today's ISO live session starts on vbox \o/ thank you all [08:46] cc xnox ^ [08:46] good [08:46] sick blud [08:49] * didrocks wants some secondary/third opinions on a dist-upgrade potential issue [08:50] so, end of last week/start of this week, I was able to reliably trigger this state during dist-ugprade (and once done): https://i.imgur.com/rgEy56h.png [08:50] meaning, GNOME Shell unloading icons, missing entries in the menu and so on [08:50] (until reboot on cosmic) [08:50] but it seems now that the package order/(low memory related? (I was still at 1Gb)) changed [08:50] and I can't reproduce it anymore [08:51] so unsure if it was the Shell going OOM or package install order triggering a Shell reloading issue [08:51] there was a way to workaround: asking the Shell to restart after dist-upgrade [08:51] I even have a do-release-upgrade Quirks for this: http://paste.ubuntu.com/p/jV7sXFgknq/ [08:51] but as ubuntu.css disappears, it's loading default Shell theme (until reboot, again) [08:52] and ofc, only works on Xorg [08:52] unsure if that worthes pushing the patch (worked, I tested it) [08:59] didrocks, I read what you wrote but I don't grasp enough what we are talking about to have an opinion [08:59] is that cosmic to cosmic updates? [08:59] or bionic to cosmic? [08:59] bionic to cosmic [09:00] that's pre reboot g-s reload, so it would mean reloading the 3.28 from the bionic session so it picks up yaru dynamically before rebooting? [09:00] no [09:00] so as you can see in https://i.imgur.com/rgEy56h.png [09:00] we could end up in a situation where there are some icons missing [09:00] no more background [09:01] and menu entries empty [09:01] (that's what I can't reproduce anymore) [09:01] one way to workaround it, is to ask G-S to restart after the dist-upgrade [09:01] but that loads then default Shell theme (upstream one) [09:01] as there is no more ubuntu.css and no pointer to Yaru ofc [09:02] so, I have a patch ready for this [09:02] but as I can't reproduce it now (seems like the package install order changed since I found issue) [09:02] but you didn't have enough record about the missing icons issue to know what it was? [09:02] like .desktop renamed? [09:02] I don't know if that worthes it [09:02] or g-s js being screwed? [09:02] sounds like it's only g-s being screwed as it's not only desktop icons that are missing [09:03] (and blue background and missing entries…) [09:03] difficult to say if it's worth it [09:03] we got no report so far so I wouldn't say it's high importance [09:03] if it's a side effect of oom I wouldn't bother [09:03] no report -> because people don't dist-upgrade yet IMHO [09:03] but yeah, no data :( [09:03] anyway, we have a patch ready [09:03] I'd leave it too IMHO [09:03] let's keep it somewhere [09:03] and revisit [09:03] k, good [09:04] sounds good [09:04] thx [09:04] just wanted to have your opinions on it as well :) [09:04] good one to raise [09:04] ok, we are all in agreement, good ;) [09:04] :) [09:04] the patch is tested, so if we really need it, it should be quick to push [09:05] * didrocks back on gdm now [09:12] good luck with gdm [09:12] I'm going to do some dist-upgrades tests in the afternoon [09:13] jibel, did you test bionic->cosmic upgrades much? if so what's the status for you, did you ever see the bug Didier was just describing? [09:14] seb128, not thoroughly. Mostly standard upgrades [09:14] went fine? [09:14] yes [09:14] what is the bug? [09:14] same for me [09:14] didrocks described it just above & shared a screenshot [09:15] jibel, see recent backlog, basically https://i.imgur.com/rgEy56h.png [09:16] never seen this or a report of a similar issue [09:18] jibel, FWIW - Hyper V works fine [09:19] great [09:19] hum, gjs_dumpstack() doesn't give anything… :/ [09:22] didrocks, can you report a bug with ubuntu-bug ubuntu-release-upgrader ? it's easier to help and try to understand what happened during the upgrade than reading lines of chat on irc [09:24] jibel: sure (even if the bug isn't really in ubuntu-release-upgrader) [09:25] let me finish current gdm inspection first [10:00] willcooke, i've changed my mind... i think i hate vbox now, Laney can we back all of those changes out?! =) [10:03] ❌ [10:08] Laney, this is not Britain's got talent! [10:10] jibel: bug #1797353 [10:10] bug 1797353 in ubuntu-release-upgrader (Ubuntu) "bionic -> cosmic dist-upgrade ends up with graphical Shell artefacts" [Undecided,New] https://launchpad.net/bugs/1797353 [10:10] xnox: heh [10:10] is there an actual problem? [10:10] nah, banter. [10:11] i do not like the "FOO is ready" notifications that appear in gnome-shell [10:11] but i don't know if it's gnome-shell problem, or my upgraded systems, or individual apps [10:11] focus stealing prevention stuff [10:11] doesn't work as well as we might like [10:12] on the list for next cycle [10:12] erm [10:12] well, we talked about it at least [10:13] it is definitely well known [10:21] Ahaha seb128 ¦ willcooke, morning 😅 [10:22] good afterning Trevinho [10:23] hey Trevinho [10:23] Trevinho, you can't have had enough sleep yet [10:23] go back to bed [10:30] Naaa... I woke up without alarm. Don't need that 😅 [10:32] saw the logs. going to study them [10:34] hey Trevinho [10:34] having fun? [10:47] Laney: ciao Iano Lano! [10:47] Mh kind of... [10:47] Mark Trevor in the house [10:53] lol === ecloud_wfh is now known as ecloud [11:32] 😂 [11:32] willcooke: do you also have the full logs for the broken and working logins? [11:33] I can get them [11:33] Trevinho, you want with all the gdm extra debugging and plymouth debugging etc etc? [11:33] and do you want me to apply laneys patch and try and recreate [11:33] I should probably do that first, since you dont want to be looking for a problem which is already solved [11:33] willcooke: full boot's journalctl , to have some context on those. [11:34] k, stand by [11:34] as the logs you sent me in both working and unworking mode the tool is actually working fine [11:34] and so should do mutter, but it doesn't for some reason.... [11:34] a very race racey [11:35] racey race racey [11:36] Trevinho, here is a working boot: https://paste.ubuntu.com/p/BjGNP7dhQd/ [11:36] racey mcraceface [11:37] that one worked?!?!?!?!?!?!?!?! [11:37] Oct 11 12:24:42 test-Inspiron-3137 gnome-shell[742]: Failed to create backend: Could not find a primary drm kms device [11:37] oh [11:38] I probably "fixed" it and logged in at some point this morning [11:38] ok, that can be a broken one [11:39] trying to get a working boot now [11:39] Laney: that could go through X then [11:39] why does it ever fail if X fallback works? [11:40] your bug, I'll step away :P [11:41] eh, that's the enigma [11:41] Even because the tool I hacked (based on what mutter does) always seem to find the very same thing: [11:41] Oct 11 12:24:46 test-Inspiron-3137 gnome-gpu-paths.desktop[1047]: Found GPUs 0x5640a64dd200, size 1 [11:42] mutter build with some debugging in that get_gpu_paths function? [11:43] yeah, I wanted to avoid that so I recreated it separately, but that seem to work all the times. [11:43] but there's probably no other way now [11:43] :< [11:44] i'm also quite sure that adding some printing makes things slow enough to make it happen less xD [11:45] question is also why X doesn't work in this case [11:46] needs drm device too no? [11:46] yeah, but from logging it seems to find it [11:50] Trevinho, working boot: https://paste.ubuntu.com/p/v3zZGBMJqv/ [11:54] Trevinho: do you mind adding this to the next gnome-shell release https://gitlab.gnome.org/GNOME/gnome-shell/commit/c17ba90209efa31b63df7c64a373c093eb33cca5 ? [11:55] ACK [11:55] it has already been merged in master and 3.30 branches but not in the 3.30.1 release [11:55] also it needs to be cherry-picked in bionic too [11:56] Trevinho: https://paste.ubuntu.com/p/XmMNstnJ2m/ <- that was a failed one from last night [11:56] it doesn't have your tool in it, but [11:56] Oct 10 17:34:34 test-Inspiron-3137 gdm-launch-environment][807]: pam_systemd(gdm-launch-environment:session): Failed to create session: Start job for unit user-123.slice failed with 'canceled' [11:57] well: [11:57] Oct 10 17:34:34 test-Inspiron-3137 systemd-logind[466]: New session c2 of user gdm. [11:57] Oct 10 17:34:34 test-Inspiron-3137 systemd[640]: Stopped Pending report trigger for Ubuntu Report. [11:57] Oct 10 17:34:34 test-Inspiron-3137 systemd[640]: Closed D-Bus User Message Bus Socket. [11:57] Oct 10 17:34:34 test-Inspiron-3137 systemd[640]: Reached target Shutdown. [11:57] Oct 10 17:34:34 test-Inspiron-3137 systemd[640]: Starting Exit the Session... [11:57] Oct 10 17:34:34 test-Inspiron-3137 gdm-launch-environment][807]: pam_systemd(gdm-launch-environment:session): Failed to create session: Start job for unit user-123.slice failed with 'canceled' [11:57] and if you look at https://paste.ubuntu.com/p/v3zZGBMJqv/ good one, the order is different === pitti is now known as cockpitmeister === cockpitmeister is now known as pitti [12:57] didrocks: are you working on a bug where you can boot to a login screen but once you login it freezes and you need a hard reset ? [13:02] bcurtiswx: does that happen every time? [13:02] jbicha: hi! yes it does [13:03] i can go into recovery and it works as expected [13:03] often that's because you have some gnome-shell extensions that aren't compatible [13:04] interesting, i don't recall adding any. Is there a way to purge all them? [13:04] try moving ~/.local/share/gnome-shell/ somewhere else and see if that helps you log in [13:05] sure, brb [13:08] jbicha: nope [13:09] ty though [13:10] bcurtiswx, does it go away if you downgrade gdm to 3.30? do you use non-wayland either by config or by using the nvidia binary drivers? [13:11] seb128: i know it's NVIDIA, and i was the same with 3.30 iirc [13:11] it* [13:11] k, not the same as Didier then [13:11] his in new in .1 [13:12] gdm using 100% cpu and making things very sloooow [13:12] i'll go test 3.30 just in case my brain isn't right. is there a quick command for that, im still a bit rusty [13:13] ah google found it :) [13:28] hmm maybe not as easy as that.. 3.30 isn't available [13:28] bcurtiswx, k, pro typehttps://launchpad.net/ubuntu/+source/gdm3/+changelog [13:28] https://launchpad.net/ubuntu/+source/gdm3/+changelog [13:29] click the version you want, e.g https://launchpad.net/ubuntu/+source/gdm3/+changelog [13:29] then click the arch, e.g amd64 [13:29] then [13:29] wget https://launchpad.net/ubuntu/+source/gdm3/3.30.0-0ubuntu2/+build/15337152/+files/libgdm1_3.30.0-0ubuntu2_amd64.deb https://launchpad.net/ubuntu/+source/gdm3/3.30.0-0ubuntu2/+build/15337152/+files/gdm3_3.30.0-0ubuntu2_amd64.deb https://launchpad.net/ubuntu/+source/gdm3/3.30.0-0ubuntu2/+build/15337152/+files/gir1.2-gdm-1.0_3.30.0-0ubuntu2_amd64.deb [13:30] then dpkg -iO *.deb [13:30] well, assuming you don't dev for GNOME and don't have the -dev installed [13:30] or you want to wget that as well [13:30] OK, so just wget everything and dpkg -iO it all [13:30] ty [13:32] wget -r -np --accept '*deb' [13:33] if you don't mind downloading them all, less copy and pasting :-) [13:35] well seb128 it is just 3.30.1 as i didn't freeze this time [13:35] i can provide didrocks with logs if he wants them [13:35] k, good! [13:35] didrocks, you are not alone :) [13:36] try installing 3.30.1 and removing /lib/udev/rules.d/61-gdm.rules [13:36] if that fixes it, probably the same thing [13:36] ok [13:40] assuming you didn't set WaylandEnabled=false in /etc/gdm3/custom.conf that is [13:42] Laney: confirmed, same thing [13:42] grand [13:42] you guys should find out if you have the same card [13:42] or if it's more widespread [13:43] using 390.87 with a GeForce GTX 970 [14:02] Laney, http://people.canonical.com/~ubuntu-archive/component-mismatches-proposed.html , britney doesn't seem happy with the provide? :/ (well it picks up on libnotify, but same situation) [14:09] seb128: I think the underlying problem is that germinate doesn't handle that situation very well [14:09] it was a pain to get xterm out of main [14:10] notification-daemon is still on today's cosmic iso [14:11] right, I demoted it this morning to see [14:11] L_aney though the gnome-shell Provides might be enough [14:12] I wonder if making libnotify recommends "notification-daemon | gnome-shell" would be enough to make component mismatch not complain (and if it's the right change) [14:13] I had to add gnome-terminal first in the list for http://ubuntudiff.debian.net/?query=-FPackage+xinit [14:14] germinate didn't seem to care that we were already seeding gnome-terminal :( [14:15] there is no actual problem though, it's only the reporting which looks buggy [14:15] which the r-t might not like though... [14:18] yeah, demoting xterm was important because I didn't want its .desktop cluttering the Activities Overview [14:19] well, notification-daemon is demoted now, it's just reported on component-mismatch (wrongly) [14:19] so technically there is no archive problem, just buggy reporting [14:25] it's on component mismatches because n-d is still on the iso [14:26] or maybe component-mismatches and germinate do about the same wrong thing :) [14:28] I think you have it backward [14:29] Source and binary movements to main (unsubscribed) [14:29] notification-daemon: notification-daemon [14:29] [Reverse-Recommends: libnotify4 (MAIN)] [14:29] it's listed because libnotify is on the ISO/main [14:29] and recommends n-d [14:29] and component-mismatch doesn't understand that a provides matches that requirement [14:29] or at least I think that's the pb? [14:32] Trevinho, Laney - dont know if this is related, or just me, but I've just noticed that when i get a working session, I can't switch vts [14:36] willcooke, did you try to configure a vpn on a new install to confirm the bug from cyphermox? [14:36] cyphermox, did you manage to report the pb upstream btw? [14:38] yes, and they provided a patch [14:38] Laney says it works, here it doesn't seem to [14:38] :( [14:41] seb128, cyphermox was this openvpn? [14:41] or any [14:45] cyphermox, do you have the reference to the bug/fix? [14:46] Laney, want me to backport a fix so you can be the one reviewing in the queue? [14:48] seb128, cyphermox have just connected to a PPTP VPN [14:48] not pptp, openvpn [14:49] bug #? [14:50] willcooke, https://bugs.launchpad.net/ubuntu/+source/network-manager-openvpn/+bug/1797236 [14:50] Ubuntu bug 1797236 in network-manager-openvpn (Ubuntu) "network-manager openvpn settings not being saved" [Undecided,New] [14:50] although DNS now seems broken [14:50] but I can't find the upstream reference [14:50] not the discussion on the IRC log of that channel or devel/r-t [14:50] shrug, why are those info in the bug :( [14:50] not->nor [14:59] Trevinho, correction - I can switch vts but its very very very slow [15:00] and gnome shell is using 100% CPUO [15:00] *CPU [15:00] and xwayland, which is odd [15:02] willcooke, well, unping for now about VPN until cyphermox or Laney can shared the bug/commit reference, unsure where they had the discussion I can't find a record [15:03] seb128, I'm trying an openvon now anyway [15:03] thx [15:03] I'm also concerned that when I disconnected the PPTP VPN I lost DNS [15:03] ffs [15:04] jibel, could you help test the vpn situation too? ^ [15:04] I can give you PPTP creds if you need some [15:04] willcooke, sure [15:05] merci [15:06] https://gitlab.gnome.org/GNOME/NetworkManager-openvpn/commit/a41d93ee654cb77ba117c648006abce5654a053c [15:06] ^ but that doesn't allow me to correctly save TLS Auth properties still; so some VPNs just can't work at all [15:07] Laney reported it was working for him [15:07] so *shrugs* [15:09] cyphermox, are you using the Canonical VPN as an example? [15:09] yes [15:09] hrm [15:09] I think Till had a problem with that too [15:09] I'll email him [15:09] cyphermox, thx [15:10] willcooke, the pb from Till was another one which we fixed and he confirmed the fix was working [15:10] it was a button not getting enabled in g-c-c [15:10] ahhh [15:10] right [15:12] well, we should get that fix ^ anyway, it's a step in the right direction [15:12] yeah [15:12] * Laney shrugs [15:12] if it's buggy seems a bit pointless [15:13] that bug could equally be SRUed [15:14] Laney, Mahtieu said it was working for you [15:14] but not him [15:14] so it's not pointless, if it makes things work for some :) [15:15] as you wish, you have my opinion on it [15:15] we install that plugin by default started this cycle, it feels like a fail if it's not working until you install SRUs [15:15] I don't agree that taking time to SRU properly working fixes is any kind of failure at all [15:15] or we should remove the mention from the release note/hope reviewers don't test that feature [15:15] but I think I say something like this every cycle [15:16] I'm not arguing that we should land a workaround or bad fix [15:16] but if upstream has an obvious invalid-use fix it's still early enough to include it imho [15:17] I don't like memory corruptions, and it's still good to have one less [15:17] even if that's not the problem Mathieu is having [15:17] anyway [15:17] that's my position [15:17] discussed enough, let's move back to fixing things [15:18] thx for the input Laney [15:18] cyphermox, how can I debug a DNS issue. When I disconnect the PPTP VPN DNS stops working. My guess is that it's still looking at the VPN's DNS [15:18] willcooke: networkctl output, pretty much [15:19] networkctl says its not running [15:20] nmcli shows the dns as the correct one on my LAN [15:25] what is it on the local machine which is doing DNS lookups? [15:25] like, resolve conf points at 127.0.0.1 [15:33] systemd-resolve --status shows the correct dns server [15:38] bcurtiswx: 01:00.0 VGA compatible controller: NVIDIA Corporation GP107M [GeForce GTX 1050 Mobile] (rev a1) [15:38] so not the same [15:39] Laney: do you plan to do a gdm upload? I remember you have a rls bug against it, correct? [15:40] just to know if I just stage the change or upload as well [15:41] didrocks: no I don't have anything in plymouth [15:41] in gdm*, it was in plymouth [15:41] Laney: ok, I'll upload right away then [15:42] makes sense [15:42] thx ;) [15:49] OK didrocks. If I can help with any logs, let me know. [15:50] bcurtiswx: right now, we are just going (in aggreement with upstream) removing the nvidia from wayland blacklist [15:50] as I can reproduce it here easily (unfortunately :p) I don't think I'll need more logs, but will do if anything needed! [15:50] 👍 [15:54] bcurtiswx: do you mind just filing your nvidia card specs on bug #1797355? [15:54] bug 1797355 in mutter (Ubuntu) "Once logged in, GDM Shell process CPU usage spiking at 100%" [Undecided,New] https://launchpad.net/bugs/1797355 [15:58] after my lunch i'd be happy to, bbl [15:58] kenvandine, hi, https://launchpad.net/~gnome3-team/+snap/gnome-calculator [16:13] ricotz: I heard jbicha bumped the gtksourceview requirement. I'm traveling on holiday now without a computer, so can't fix it this week. [16:16] seb128, FYI: https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1797415 [16:16] Ubuntu bug 1797415 in systemd (Ubuntu) "DNS stops working when disconnecting PPTP VPN on desktop" [Undecided,New] [16:22] kenvandine, I see, enjoy your holiday [16:27] ricotz: thanks [16:27] kenvandine, can I help with the gnome-calculator issue? [16:28] looking at what the build failure is [16:28] oSoMoN: probably, I think a backport needs to be updated [16:29] oSoMoN, jbicha bumped a build requirement in master [16:30] The failure is in our automatic build of our git mirror [16:30] my guess is that you just need to build cosmic's gtksourceview4 in your ppa [16:30] oSoMoN: actually, just turn off the automatic building [16:30] oSoMoN: ^ [16:30] good idea :) [16:31] We never promote that version anyway [16:31] We only promote the tagged versions [16:31] ack, I turned off the automatic builds [16:31] Thanks [16:32] shall I try backporting gtksourceview4 to xenial? [16:32] I'll need to bump that in the PPA before 3.32 [16:32] oSoMoN, nah... Not right now [16:32] ok [16:32] Build snap will solve it :-) [16:33] Thanks, I'm going back to being on holiday :-) [16:35] kenvandine, enjoy! === pstolowski is now known as pstolowski|afk [17:08] willcooke, k [17:14] huh, it got late [17:14] :) [17:17] guess cosmic is finished tho, happy final freeze ;-) [17:17] heh [17:17] well not quite, 2h45 left [17:18] hear that Trevinho? [17:18] ;D [17:32] yeah :P [17:35] Trevinho, those extra logs any use? [17:37] willcooke: mh, still trying to figure out a pattern :/ [17:38] Trevinho, did you see my comment about not be able to switch vts earlier? Looks like xwayland was crashing(?!?) that cause the slow down [17:38] yes... [17:38] that's when it boots fine, right? [17:38] let me test is again [17:39] ok, booted in to a working gdm [17:39] is it x11 or wayland? [17:39] as on x11 ofc there shouln't be that [17:40] `sudo strings /proc/$(pidof gnome-shell)/environ | grep XDG_SESSION_TYPE` from ssh should give you that [17:41] k, thats in Wayland it seems [17:41] now I will try and switch vt [17:41] yup - I can't [17:42] however, gdm is still responsive [17:42] so maybe that's a feature? [17:42] in theory it should stop it and start again on switch [17:42] logged in ok [17:43] switching back to 1 [17:43] it starts [17:43] but I can use the usual trick to find out if its x or wayland [17:43] sudo strings /proc/$(pidof gnome-shell)/environ | grep XDG_SESSION_TYPE [17:43] strings: Warning: '/proc/2190' is a directory [17:43] strings: '1739/environ': No such file [17:44] looking at ps [17:44] ah, lloks there are two isntances running [17:44] get the id from ps [17:45] although is weird [17:45] https://paste.ubuntu.com/p/2NZPP49f52/ [17:45] thats the PS output ^ [17:45] I can switch vts fine [17:45] and no xwayland crash [17:45] `2190` should be there [17:45] so that might have been a red herring [17:46] is it respawning, causing there to appear to be two ? [17:46] $ sudo strings /proc/$(pidof gnome-shell)/environ | grep XDG_SESSION_TYPE [17:46] XDG_SESSION_TYPE=x11 [17:46] mh ok [17:46] ah [17:46] it's when I switch to vt1, then vt2 is still running [17:46] where my session is [17:47] when I switch back to 2, the proper session, and gdm shuts down again... [17:47] then the strings thing works [17:47] for the two failed logs you sent me (both x11 and wayland) you didn't save the journalctl, right? [17:48] as you sent me the procs, but not journalctl so not sure what was making it fail [17:48] ok, let me fix up a script to do this for me [17:48] and then I will send you a few good and bad logs [17:49] good... add both envs and stuff, but even full journalctl please [17:49] "both envs"? [17:50] sudo strings /proc/$(pgrep -U gdm -x gnome-shell)/environ | grep XDG_SESSION_TYPE [17:50] ah right === tomreyn_ is now known as tomreyn [17:56] let me also try if a clean installation here triggers it... [17:56] that pgrep -U gdm ..... isnt returning anything [17:56] I could just look at the pid for gdm3 right? [17:57] oh, ignore [17:57] I undersatnd now [17:58] should be the same that just filter on user, in case there are multiple instances [17:58] but shouldn't be an issue in this case as there is only one [18:03] Trevinho, ok, script is working [18:03] I'll do a few boots until I catch at least 3 broken ones [18:04] k good thanks [18:12] mh, my udev seems fast enough, can't reproduce this... :( [18:13] Suddenly I cant either [18:13] :) [18:16] ok, I'm going to give it the magic "10 more boots and I'll declare it fixed" thing [18:18] ok, something different is going on now [18:18] before I would get the cursor and that was an indication that it was not goiong to work [18:18] now I get the cursor, and a few seconds later the mouse appears and it works [18:22] yeah, so as normal... [18:22] while before was like blinking right? [18:22] yeah exactly [18:22] as per the retries.. [18:23] (I think I might have done an upgrade on this box earlier,.... ) [18:23] laney's patch is already in? [18:23] I wonder [18:24] as one thing I noticed was that in the previous logging there where cases where gdm was starting gnome-session when logind wasn't grahical yet. While we should wait for that. [18:26] I'm going to try silly stuff now, like booting without the ethernet cable in [18:30] going offline for a long week-end, happy final freeze and have a good week-end desktoppers! [18:31] Trevinho, right I'm 20 for no fails [18:32] I think it might be EOD [18:32] Because tomorrow it will break again [18:41] Laney, thx for the vpn upload, sorry about the assignment, cyphermo_x said here that you had tried a patch was resolving the issue for you but not him so I though you had a fix that was confirmed to be working for the issue you reproduced [18:45] question: why is the version mentioned in that patch 1.8, when we're shippnig 1.12 [18:45] oh [18:45] becuase it's the openvpn bit that is version 1.8 not n-m [18:45] thanks me [18:46] I am tired [18:46] Right. vbox is fixed. This gdm thing I now can recreate [18:46] I've logged a bug for the pptp dns thing [18:47] the n-m security issue from last night is all OK [18:47] The screenshots for the new hyper v image are done [18:48] firefox esr notice is going out tomorrow [18:48] willcooke: well I really hope it was related to the fact that some pieces where not updated, as a part than waiting what laney already did I wouldn't have much other idea [18:48] willcooke: monitor what the scripts output in case [18:49] Trevinho, if it's just me left that can sometimes reproduce it, I think we should probably give up. [18:49] There are more important things for you to spend your time on than chasing a maybe imaginary bug [18:49] I'll keep trying [18:49] but not today [18:49] not imaginary for sure, just quite hard to debug seen the fact there's not a clear pattern, other than waiting udev, as should be done now [18:50] edit: Right. vbox is fixed. This gdm thing I now can recreate -< *CAN NOT* [18:52] Remember I was asking about monitors back in July. [18:52] The one I wanted, the Dell U2414H, was 160 quid, now 205 quid. [18:52] gr [18:53] I?ve to leave, dinner time. I guess we have still work for quick SRUs [18:53] see you Trevinho [19:00] that'll do for today. night all [19:03] seb128: no worries, I figured after all since you made me think about it again that I'd try to reproduce in a VM :P [19:04] recording screencasts is quite fun [19:08] Trevinho: that CanGraphical thing is a red herring, stop thinking about that [19:08] the plymouth fix is indeed [19:08] in [19:09] well, you can verify CanGraphical by putting loginctl show-seat seat0 in ExecStartPre but we already had lots of cases where this was 'yes' and the bug still happened [19:09] I think we got excited because it seemed to be in a similar area, but it's actually not what you're looking for here [19:14] Laney: is wrong to set is as start pre. Has to be launched by gnome session. Anyway, will has logs with that (my tool was checking that too). I can show them later [21:57] Trevinho: why? [21:57] you're speculating that it goes yes -> no in that tiny window? [21:59] Laney: see this log https://paste.ubuntu.com/p/j4PMxt24K6/ [21:59] grep for gpu-paths.desktop[ [22:02] for checking gpu we need to have the systemd session name up, and use it for taking control if needed, but well... as you can see sometimes it was not ready. that's why wayland not always work. and sometimes the x11 fallback works otherwise doesn't. [22:02] what happened is sometimes uses wayland sometimes x11. sometimes none of them. [22:05] https://paste.ubuntu.com/p/p7vGPdVndC/ [22:05] there's a log from the MR [22:07] Oct 08 13:55:32 test-Inspiron-3137 gdm3[591]: GdmLocalDisplayFactory: creating seat proxy for seat 'seat0' with path '/org/freedesktop/login1/seat/seat0' [22:07] Oct 08 13:55:32 test-Inspiron-3137 gdm3[591]: GdmLocalDisplayFactory: wayland login display for seat seat0 requested [22:07] that means that sd_seat_can_graphical() returned TRUE [22:07] then: Oct 08 13:55:33 test-Inspiron-3137 gnome-shell[739]: Failed to create backend: Could not find a primary drm kms device [22:07] * Laney gtg [22:17] Laney: yeah, the g-s error is actually the same codepath of the tool [22:17] Laney: https://gist.github.com/3v1n0/3e149eb7e21f166ccee1700f313ce1dc [22:34] Hey kenvandine, you around?