| seb128 | goood morning desktopers, today I forgot to start my IRC client again! | 09:52 |
|---|---|---|
| juliank | hmm the workspace indicator that replaced activities in the top left isn't actually an upstream bit and just leaked into vanilla gnome sessions right? | 09:54 |
| seb128 | juliank, it is upstream | 10:03 |
| juliank | oh then the fedora folks are missing it yet :D | 10:12 |
| juliank | we do weird culture clash at all systems go | 10:16 |
| jbicha | yeah, Fedora 39 Beta will have the new workspace indicator when it releases (perhaps next week) | 12:18 |
| ogra | GRRRR !!! have these upstream devs that decided that modal popup windows should be locked in the middle of the main window of an app ever used spellchecking in any app ? this is hilarious, trying to spellcheck a long email in evolution i can not even scroll in the main window to see where he check is | 12:42 |
| ogra | what a silly default ... | 12:43 |
| ogra | (sorry, just ranting) | 12:43 |
| mdeslaur | ogra: I hate hate hate hate hate that | 12:47 |
| jbicha | ogra: you know how to disable that right? | 12:52 |
| nteodosio | Seems to be the same with the Find window :O | 12:53 |
| ogra | i was guessing i can .... but wasnt really after digging through dconf-editor | 12:53 |
| nteodosio | Oh so is that a GTK thing? I'm interested in disabling it then. | 12:54 |
| ogra | more a gnome shell thing | 12:55 |
| ogra | (likely) | 12:55 |
| nteodosio | I'm on dwm, it's the same behavior. | 12:55 |
| ogra | ah, interesting ... yeah, then it must be the toolkit | 12:55 |
| jbicha | ogra: install gnome-tweaks . In Tweaks, switch to Windows page, then turn off "Attach Modal Dialogs" | 12:56 |
| ogra | jbicha, well, i'm more quesioning the reason for us to follow that defaault since it is a real usability nightmare | 12:57 |
| ogra | +t | 12:57 |
| nteodosio | It looks like a bug actually. Despite I give focus to the main window and the cursor blinks there, if I type anything it goes to the transient one. | 12:57 |
| jbicha | gsettings is org.gnome.mutter attach-new-windows | 12:57 |
| jbicha | because it's gsettings we could override it in the ubuntu-settings package for the Ubuntu sessions if we wanted | 12:58 |
| ogra | thats what i mean ... | 12:58 |
| ogra | i have the feeling we just blindly folloed that change when it was introduced | 12:59 |
| jbicha | window management in default GNOME feels rough to me. There's no app menu in the top bar anymore with GNOME 45. It's a lot easier to see what app is focused in Ubuntu with our dock | 13:00 |
| jbicha | for some apps it is helpful, gnome-control-center still uses popups for things | 13:03 |
| juliank | fwiw, gnome-shell now seemingly gets killed after each lid close and reopen | 13:19 |
| juliank | I don't get any crash error messages, though | 13:19 |
| ogra | might be a feature then 😛 | 13:20 |
| juliank | heh | 13:20 |
| juliank | need to monitor and debug more | 13:21 |
| juliank | the first error logged is | 13:22 |
| juliank | Sep 13 13:16:41 jak-t14-g3 firefox-bin[48768]: Error reading events from display: Broken pipe | 13:22 |
| juliank | Sep 13 13:16:41 jak-t14-g3 systemd[2990]: org.gnome.Shell@wayland.service: Main process exited, code=killed, status=9/KILL | 13:22 |
| juliank | no reasons | 13:22 |
| juliank | no crash dumps either | 13:22 |
| ogra | that looks like an explicit kill -9 from somewhere, no ? not really like a crash ... | 13:24 |
| juliank | oh | 13:24 |
| juliank | I think I accidentally made systemd kill the user session when it locks the screen | 13:24 |
| ogra | heh | 13:24 |
| juliank | I set KillUserProcesses=yes and I guess due to homed-related changes this also applies to suspend now | 13:25 |
| juliank | I had to do a lot of login/logout testing and needed to make sure I got clean sessions | 13:25 |
| seb128 | juliank, https://bugs.launchpad.net/ubuntu/+bug/2034619 | 13:32 |
| -ubottu:#ubuntu-desktop- Launchpad bug 2034619 in Ubuntu "gnome-shell gets SIGKILL'd when lock screen in Wayland" [Undecided, Confirmed] | 13:32 | |
| seb128 | juliank, Daniel was trying to get the submitter to provide more information, maybe you can help there | 13:32 |
| juliank | seb128: Right it seems crashing reproducibly, even after just removing the logind.conf line | 13:34 |
| juliank | I press Windows+L key and get to the login screen | 13:34 |
| === pizzaiolo is now known as pizza | ||
| juliank | only ubuntu appindicators enabled | 13:35 |
| juliank | not sure how to debug the source of the kill signal | 13:41 |
| juliank | jak@jak-t14-g3:~:master$ sudo strace -p $(pidof gnome-shell) -e 'trace=!all' | 13:46 |
| juliank | strace: Process 33093 attached | 13:46 |
| juliank | +++ killed by SIGKILL +++ | 13:46 |
| juliank | thanks for nothing | 13:46 |
| juliank | gdb clean too | 13:46 |
| juliank | I guess I could somehow use bpf to trace this in the kernel | 13:47 |
| seb128 | juliank, could you perhaps try what Daniel suggested and see if Xwayland is crashing or something? | 13:51 |
| juliank | I can attach to Xwayland I suppose | 13:51 |
| juliank | seb128: there was no Xwayland running even | 13:52 |
| seb128 | k, so probably not that... | 13:52 |
| juliank | I'm asking the web hivemind | 13:54 |
| juliank | how to log kill signals in the kernel | 13:54 |
| seb128 | juliank, https://access.redhat.com/solutions/165993 suggests maybe 'audit' can help? | 13:55 |
| seb128 | juliank, or perhaps killsnoop | 13:56 |
| juliank | OK I trapped a lot of kills | 13:58 |
| juliank | it seems dbus-daemon is the only one sending kill signals | 14:04 |
| juliank | but I don't see a kill signal to the pid of gnome-shell, maybe it's sending it to a different thread pid | 14:04 |
| juliank | Gotta go do talk prep but leave that here | 14:10 |
| juliank | jak@jak-t14-g3:~:master$ sudo killsnoop-bpfcc -T $(pidof gnome-shell) | 14:10 |
| juliank | TIME PID COMM SIG TPID RESULT | 14:10 |
| juliank | 16:09:39 711 systemd-journal 0 99853 -3 | 14:10 |
| juliank | 16:09:43 115187 gnome-shell 0 99853 -3 | 14:10 |
| juliank | 16:09:43 115187 gnome-shell 0 99853 -3 | 14:10 |
| juliank | oh but they are all SIG 0 | 14:10 |
| juliank | sigh | 14:10 |
| juliank | maybe lennart has a better idea | 14:12 |
| juliank | I bpf traced all SIGKILL and only saw dbus-daemon either | 15:09 |
| juliank | it's possible we're getting killed *by the kernel* and not user space and need to ftrace do_exit | 15:09 |
| Eickmeyer | juliank: I had the same issue, but in X under Nvidia, so I don't thinkit's isolated to just wayland. | 15:21 |
| Eickmeyer | I think you're on the right track. | 15:21 |
| juliank | It's complicated, I don't know how to ftrace yet | 15:21 |
| juliank | I am at All Systems Go so I should be able to find _someone_ to hlep me | 15:22 |
| Eickmeyer | Mine gave me a forced logout. | 15:22 |
| juliank | I downgraded dbus because I saw sigkills from dbus but that didn't help but maybe I didn't go back far enough | 15:22 |
| Eickmeyer | I'd say bisect but it's a little late. | 15:24 |
| juliank | I'll try to talk to brauner and get some help with fatrace | 15:36 |
| juliank | * ftrace | 15:36 |
| Eickmeyer | seb128: Hi! Before I forget, an upload of ubiquity-slideshow-ubuntu would be nice. I had an installability bug in a couple of other packages to fix, so I (sadly) didn't look for another sponsor. | 15:41 |
| Eickmeyer | Thought it should be something to get in before beta freeze on Monday. | 15:43 |
| seb128 | Eickmeyer, oh, thanks for the reminder, I forgot on Monday | 15:59 |
| seb128 | I said I would do it after the weekend | 16:00 |
| Eickmeyer | :) | 16:01 |
| ahasenack | hi #desktop, I downloaded the mandic daily iso to try out the installer with TPM support for FDE, but I don't see the option, is it published yet? | 18:13 |
| ahasenack | or do we need a special flag to enable it somehow? | 18:13 |
| ahasenack | ref: https://ubuntu.com/blog/tpm-backed-full-disk-encryption-is-coming-to-ubuntu | 18:14 |
| ricotz | seb128, hello :), libreoffice amd64 waits in the NEW queue | 18:25 |
| ricotz | seb128, can you accept libreoffice amd64 from the NEW queue? | 20:12 |
| seb128 | ricotz, hey, sorry yes, you pinged while I was away for dinner and you were not the only one, getting there! | 20:13 |
| seb128 | Eickmeyer, ubuntu-slideshow-ubuntu changes uploaded | 20:13 |
| Eickmeyer | seb128: ack, thanks! | 20:14 |
| ricotz | seb128, alright, I assumed you missed it since you already answered twice in other occasions ;) | 20:15 |
| seb128 | ricotz, accepted | 20:19 |
| ricotz | seb128, thanks | 20:22 |
| seb128 | np! | 20:23 |
| seb128 | ricotz, I've also uploaded thunderbird 102.15.1 to the security ppa | 20:23 |
| seb128 | ricotz, and we are back to 59 riscv builders :) | 20:25 |
| KGB-2 | gnome-initial-setup signed tags d8cc56b Jeremy Bicha ubuntu/45_beta-2ubuntu1 * gnome-initial-setup Debian release 45~beta-2ubuntu1 * https://deb.li/3L2lE | 20:25 |
| KGB-2 | gnome-initial-setup ubuntu/latest fddc8ed Jeremy Bicha * pushed 6 commits (first 5 follow) * https://deb.li/Tngl | 20:25 |
| KGB-2 | gnome-initial-setup ubuntu/latest f57acad Amin Bandali debian/gbp.conf * Change debian-branch from debian/master to debian/latest * https://deb.li/IpKd | 20:25 |
| KGB-2 | gnome-initial-setup ubuntu/latest b3a94eb Jeremy Bicha debian/patches/ series Revert-keyboard-Spawn-tecla-to-show-keyboard-map.patch * Drop patch reverting use of Tecla for Keyboard panel * https://deb.li/3QYu1 | 20:25 |
| KGB-2 | gnome-initial-setup ubuntu/latest 46a772d Jeremy Bicha debian/ control control.in * Depend on tecla * https://deb.li/1Smd | 20:25 |
| KGB-2 | gnome-initial-setup ubuntu/latest 9856b22 Jeremy Bicha debian/changelog * releasing package gnome-initial-setup version 45~beta-2 * https://deb.li/3VsQV | 20:25 |
| KGB-2 | gnome-initial-setup ubuntu/latest c0e7af0 Jeremy Bicha debian/ (5 files in 2 dirs) * Merge tag 'debian/45_beta-2' into ubuntu/latest * https://deb.li/iiFII | 20:25 |
| ricotz | seb128, I noticed 102.15.1 :), don't forget about 115.2.2, tarball in ppa and pushed to git | 20:26 |
| ricotz | seb128, https://launchpad.net/ubuntu/+source/llvm-toolchain-16/1:16.0.6-15/+build/26703138 | 20:26 |
| seb128 | ricotz, I will do 115.2.2 tomorrow | 20:26 |
| ricotz | seb128, ack | 20:27 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!