[05:14] <didrocks> good morning
[05:18] <jibel> good morning everyone
[05:19] <jibel> Salut didrocks, ça va? pas trop en train de fondre?
[05:19] <duflu> Morning didrocks and jibel 
[05:19] <duflu> Also, back later
[05:19] <jibel> hi duflu 
[05:22] <didrocks> jibel: salut ! le matin à 5h, ça va :) ça commence à chauffer là, mais bon, aujourd’hui, c’est un peu le répit avant jeudi/vendredi/samedi
[05:22] <didrocks> et toi ?
[05:22] <didrocks> hey duflu 
[05:23] <jibel> didrocks, ça va bien. Content de retrouver la fraicheur bretonne.
[05:42] <oSoMoN> good morning desktoppers
[05:59] <didrocks> salut oSoMoN 
[06:01] <oSoMoN> salut didrocks 
[06:25] <marcustomlinson> morning desktoppers
[06:29] <callmepk> hi jibel duflu didrocks oSoMoN marcustomlinson 
[06:32] <didrocks> hey callmepk! Congrats again :)
[06:34] <callmepk> didrocks, thanks! that fellowship is really a surprise to me
[06:39] <jamesh> congratulations callmepk 
[06:40] <duflu> Morning oSoMoN, marcustomlinson 
[06:40] <callmepk> Hi jamesh, thanks!
[06:40] <duflu> Afternoon and congrats callmepk 
[06:40] <duflu> and afternoon jamesh 
[06:43] <callmepk> Hi duflu , and thanks!
[06:44] <oSoMoN> good morning marcustomlinson, callmepk, duflu, jamesh, jibel 
[06:45] <duflu> factorial(good morning)
[07:50] <seb128> goood morning desktopers
[07:54] <duflu> Hi seb128 
[07:58] <didrocks> salut seb128 
[08:01] <Laney> yo
[08:03] <duflu> ello Laney
[08:03] <jamesh> duflu: I've been trying to get gdm up and running on an Ubuntu Core system without much success.  I was wondering if you might have some pointers based on this log: https://paste.ubuntu.com/p/42fQkhv64y/
[08:04] <jamesh> it looks like it is having trouble switching virtual terminals, but I'm not sure which permission that indicates it's missing
[08:04] <duflu> jamesh, I know practically nothing about gdm. Probably Laney is the expert here
[08:05] <jamesh> Laney: ^^^ do you see any obvious culprits in that log?
[08:05] <duflu> jamesh, isn't the issue "VT_ACTIVATE failed: Operation not permitted" ?
[08:06] <duflu> Based on using 'chvt' I think you usually need to be root
[08:06] <duflu> or get the right privileges at least
[08:06] <jamesh> duflu: this is from doing a "sudo systemctl start gdm.service" with the standard 20.04 unit
[08:10] <Laney> I think you need CAP_SYS_TTY_CONFIG
[08:10] <Laney> anything in dmesg?
[08:10] <Laney> did you try wayland?
[08:11] <jamesh> Presumably wayland failed and it fell back to an X serssion.  I didn't change the default config
[08:12] <jamesh> I wonder if some xattrs got lost when making the image?
[08:13] <jamesh> nothing in dmesg when I try to start gdm again
[08:15] <Laney> can you turn on gdm's debugging in /etc/gdm3/custom.conf?
[08:15] <Laney> and maybe arrange for G_MESSAGES_DEBUG=all to be set
[08:15] <Laney> I'm thinking something in the wayland attempt might be less inscrutable
[08:15] <didrocks> hey Laney 
[08:15] <didrocks> seb128: do you mind subscribing desktop-packages to https://launchpad.net/ubuntu/+source/fonts-material-design-icons-iconfont ?
[08:16] <jamesh> lets see.  I can probably bind mount a new version of the config file over the top
[08:17] <Laney> moin didrocks 
[08:20] <jamesh> Laney: here's a log with debugging turned on: https://paste.ubuntu.com/p/QTTCZVfWwr/
[08:22] <seb128> lut didrocks
[08:22] <seb128> didrocks, 
[08:22] <seb128>  o fonts-material-design-icons-iconfont: fonts-material-design-icons-iconfont
[08:22] <seb128>    [Reverse-Recommends: freetype2-doc (MAIN)]
[08:22] <seb128> but freetype is owned by foundations
[08:23] <seb128> so shouldn't it be there to subscribe?
[08:23] <seb128> hey duflu, jamesh, Laney
[08:23] <didrocks> seb128: in similar cases in the past, it was desktop-team for the fonts
[08:24] <didrocks> so this is why I think we should do the same than other fonts package
[08:24] <didrocks> if you don’t want, happy to spend some time with foundations playing ping-pong
[08:24] <seb128> I think we did for things which were pulled in by one of our packages
[08:24] <seb128> not that I really care either way
[08:24] <seb128> let me try playing pingpong while you are on holidays so you don't have to be bothered by it
[08:25] <didrocks> I don’t think this is a big load for fonts package, but just tell me what is your preferred way
[08:25] <didrocks> ack
[08:25] <seb128> thx
[08:25] <didrocks> feel free to promote it once done :)
[08:25] <didrocks> (no MIR bug needed)
[08:29] <Laney> jamesh: huh, it just tried x immediately
[08:29] <Laney> is this a gdm you built?
[08:29] <jamesh> Laney: it's 20.04's gdm
[08:29] <jamesh> Laney: before it tries to start the X server, there is "gdm-launch-environment][1113]: GdmSessionWorker: could not take control of tty: Operation not permitted"
[08:29] <Laney> what's happened to it? somehow inserted into a snap?
[08:30] <Laney> I think it's going to come down to that stuff
[08:30] <Laney> apparmor does capability mediation doesn't it?
[08:30] <jamesh> I built a version of the core20 snap that includes GDM plus dependencies, then built an Ubuntu Core image that boots from that base
[08:31] <jamesh> It's just another service in the boot file system, so no special snap security policies
[08:32] <Laney> hmm
[08:35] <jamesh> Laney: I've also made /var/lib/AccountsService and /var/lib/gdm3 writable: I couldn't get as far as starting gdm without that
[08:45] <Laney> jamesh: Yeah, I'd have ~expected that
[08:45] <Laney> jamesh: try https://gist.github.com/iainlane/1277e2446e8503fc698f1a175fd4561f in ExecStartPre of the service or something?
[08:46] <Laney> I'm hoping that it's unavailable, and the question will be to figure out why
[08:46] <Laney> I don't have super loads of clue tbh
[08:46] <Laney> oh good, I got GNOME commit access back
[08:47] <jamesh> actually, I have an idea.  It might be console-conf
[08:47] <seb128> Laney, ah, you too :)
[08:48] <Laney> seb128: I was like "surely it hasnt been two years since I committed something"
[08:49] <seb128> did you at least check the calendar to see that it was not 2022 and that you just had slept for that long? :p
[08:49] <jamesh> Laney: "systemctl mask console-conf@tty1.service" did the trick
[08:49] <Laney> never heard of that
[08:49] <Laney> is it some core thing?
[08:49] <jamesh> console-conf is the first-run Ubuntu Core setup tool
[08:50] <jamesh> it looks like it was holding /dev/tty1 open, preventing GDM from starting
[08:50] <Laney> ah
[08:50] <Laney> ok, that makes perfect sense
[08:51] <jamesh> one more thing to add to my readme
[08:51] <Laney> gdm already has Conflicts/After on getty@tty1.service
[08:51] <Laney> make a drop-in to add that other one too maybe
[08:51] <Laney> unless you need to run the wizard once to get a user ...
[08:53] <jamesh> in theory I should be able to get a user account created via a system-user assertion.  Haven't got that working yet though, so I have been relying on console-conf to get a login
[08:53] <Laney> conflicting off stuff which wants to own gdm's tty is a sane thing to do in general
[08:53] <Laney> eventually you'd probably want to use gnome-initial-setup to do a similar job to the wizard
[08:54] <Laney> gdm has a "no users? run gnome-initial setup" codepath
[08:55] <jamesh> you're supposed to be able to put a removable drive with an "auto-import.assert" file containing signed instructions to create a user
[08:55] <Laney> sounds ok for a POC
[08:55] <Laney> once it works :>
[09:34] <jamesh> Laney: looks like I still have an X11 greeter session even when switching the VM to virtio-vga with virgil support.
[09:34] <jamesh> the X server logs seem to indicate it is using Virgil acceleration
[09:44] <Laney> I'd expect to see evidence of it at least *trying* wayland first
[09:44] <Laney> but the log you pasted just went straight to X
[09:44] <Laney> that's why I asked if you rebuilt it, because I was suspicious that the wayland support didn't get compiled in
[09:46] <Laney> perhaps I just forgot what the logs look like though
[09:48] <Laney> don't think virgl should preclude wayland, should it?
[09:48] <Laney> not that I've tried
[09:49] <jamesh> I was running with qxl as the vga device before: was wondering it it might be too stupid for Wayland
[09:49] <jamesh> and if virtio-vga/virgil might have more luck
[09:53] <Laney> qxl should be ok anyway: https://people.canonical.com/~laney/weird-things/qxl-wayland.png
[09:55] <Laney> if there's instructions for how to build this thing, ideally with a modified gdm, I could poke it at some point
[11:34] <Nafallo> hi guys! latest update to gnome-shell and gnome-control-center left the sound panel without any detected sound card. known bug?
[12:13] <oSoMoN> Nafallo, no other updates?
[12:29] <Nafallo> oSoMoN: not entirely sure when I restarted last time I'm afraid...
[12:30] <oSoMoN> Nafallo, I'd suggest filing a bug against gnome-control-center (using ubuntu-bug), and make sure to attach /var/log/apt/history.log
[12:30] <Nafallo> I'll corolate some logs :-)
[12:31] <seb128> check in pavucontrol first if the devices are listed
[12:31] <seb128> it seems more likely that pulseaudio isn't working
[12:31] <seb128> rather than it being a gnome-control-center bug
[12:34] <Nafallo> looking at the changelog entry for pulseaudio this looks like exactly what was trying to be solved :-D
[12:35] <Nafallo> think you might be right seb128 :-)
[12:37] <Nafallo> downgraded pulseaudio, rebooting for good measure. brb.
[12:43] <Nafallo> nafallo@wendigo:~$ systemctl status pulseaudio.service
[12:43] <Nafallo> Unit pulseaudio.service could not be found.
[12:43] <Nafallo> that looks like a problem :-P
[12:45] <Nafallo> ah --user
[13:11] <fossfreedom> seb128: re bug 1889317 ... have to admit this is a new area for me. Your comment about default.list ... are you referring to a package dpkg-divert type thing?
[13:11] <ubot5> bug 1889317 in file-roller (Ubuntu) "When clicking a .deb, file-roller gets launched ( instead of an installer for .deb )" [Low,Invalid] https://launchpad.net/bugs/1889317
[13:15] <oerheks> fileroller? that is not a bug, it is by design, AFAIK
[13:16] <seb128> fossfreedom, I don't remember offhand but I think it uses XDG_DATA_DIRS 
[13:17] <seb128> fossfreedom, so I think you could ship your own version in /usr/share/<desktop>/applications/defaults.list
[13:17] <seb128> fossfreedom, https://packages.ubuntu.com/search?searchon=contents&keywords=defaults.list&mode=&suite=focal&arch=any
[13:18] <Nafallo> seb128: thanks for the pointing, filed #1889416
[13:18] <seb128> bug #1889416
[13:19] <ubot5> bug 1889416 in pulseaudio (Ubuntu) "pulseaudio.service: Start request repeated too quickly." [Undecided,New] https://launchpad.net/bugs/1889416
[13:19] <seb128> Nafallo, could you try to start 
[13:19] <seb128> $ /usr/bin/pulseaudio --daemonize=no --log-target=journal
[13:19] <seb128> by hand and see if it prints any error?
[13:20] <fossfreedom> seb128:  Interesting.... ok definitely never played in this area. Cheers. Going to see how budgie-desktop plays here. Just hoping that our heavy gnome dependencies doesn't get in the way here
[13:21] <Nafallo> seb128: nothing at all. just starts happily.
[13:21] <seb128> fossfreedom, it shouldn't ... what's echo $XDG_DATA-DIRS on bugdie?
[13:21] <seb128> Nafallo, is there an active instance already when you try to start the unit?
[13:23] <Nafallo> seb128: well, not when I try. you mean when it fails?
[13:23] <Nafallo> the failures is when I check after reboot and login.
[13:24] <seb128> it's weird :-/
[13:24] <Nafallo> after that I can start it manually with systemctl --user start pulseaudio.service
[13:24] <seb128> did you try to downgrade to the preSRU version?
[13:24] <fossfreedom> seb128: /usr/share/budgie-desktop:/home/dad/.local/share/flatpak/exports/share:/var/lib/flatpak/exports/share:/usr/local/share:/usr/share:/var/lib/snapd/desktop:/var/lib/snapd/desktop
[13:24] <fossfreedom> ... so /usr/share/budgie-desktop/applications/defaults.list is worth a try
[13:25] <Nafallo> I can try again, just in case.
[13:25] <Nafallo> seb128: sudo dpkg -i /var/cache/apt/archives/{libpulse,pulseaudio}*_1%3a13.99.1-1ubuntu3.4_a*.deb
[13:25] <Nafallo> that one?
[13:25] <seb128> fossfreedom, right, try to just copy the /usr/share/applications/defaults.list there and edit the gnome-software line and see if it works
[13:26] <seb128> Nafallo, well, the one you had before recent updates so probably yes?
[13:26] <Nafallo> yeah, that one then :-)
[13:26] <seb128> fossfreedom, note that we do it with a file in /etc and a symlink so it's a conffile and it can be changed as such locally without being overwriting on upgrade
[13:27] <Nafallo> okay, rebooting again. brb.
[13:28] <Nafallo> seb128: this one works :-)
[13:30] <Nafallo> time to read the patch :-)
[13:38] <fossfreedom> seb128: cheers - that seems to have worked. We'll do the same with the /etc/ symlink idea.
[13:41] <Nafallo> seb128: Jul 29 14:59:18 wendigo pulseaudio[2728]: Failed to create secure directory (/home/nafallo/.config/pulse): No such file or directory
[13:42] <Nafallo> so for some reason the latest update break ext4 encrypted home directories ;-)
[13:42] <Nafallo> s/break/& on/
[13:45] <seb128> weird, the patch is basically a retry loop on an existing function
[13:48] <Nafallo> hmm. I guess the next step is to upgrade again and see if it's reproducable...
[13:52] <Nafallo> actually. looks like I got this with the lower version as well, but it kept restarting and succeeded before it gave up.
[13:53] <Nafallo> retried 4 times, and on the last attempt it started :-D
[13:59] <seb128> so is that because it tries before the userdir is decrypted?
[13:59] <Nafallo> I believe so, yeah.
[14:00] <seb128> k, so probably just not a regression, just timing and random
[14:00] <seb128> could you add the details in the bug report?
[14:00] <Nafallo> pulseaudio breaks faster, and the retry counter reaches zero before it's decrypted :-)
[14:01] <Nafallo> yeah. better do that before I try to figure out how to add a condition to the service to wait for the home directory to be available :-D
[14:01] <Nafallo> what is it with assumptions in IT? :-P
[14:05] <Nafallo> okay, updated.
=== cpaelzer_ is now known as cpaelzer
=== kedar is now known as kedar_apte
=== ijohnson is now known as ijohnson|lunch
[17:21] <luna_> Installed Ubuntu on a non broken SSD