| 19 Nov 2023 |
 Jan Tojnar | compare strings /proc/$pid/environ of Nautilus being launched by the different methods | 16:51:21 |
| Jan Tojnar | * compare strings /proc/$pid/environ of Nautilus launched by the two different methods | 16:52:19 |
 Vika (she/her) | Jan Tojnar That was my first thought! But... XDG_DATA_DIRS is almost the same between launching anything from wofi and my terminal, and both entries include /run/current-system/sw/share at the end, as they should.
| 16:54:26 |
| Vika (she/her) | In fact, almost the entire environment is the same! Sans a few variables GIO sets when launching an application using g_app_info_launch.
| 16:55:06 |
| Vika (she/her) | Specifically, these variables are XDG_ACTIVATION_TOKEN, GIO_LAUNCHED_DESKTOP_FILE{,_PID}, DESKTOP_STARTUP_ID and some minor variables like GPG_TTY (surely they don't affect this?). PATH is sometimes a bit different, but it's because of prefixes — the end is always /run/current-system/sw/share as expected.
| 16:57:49 |
| Jan Tojnar | Vika (she/her): I would try running it in terminal with XDG_DATA_DIRS=/run/current-system/sw/share and maybe some other variables cleared | 17:01:30 |
| Vika (she/her) | ...wait a second. | 17:10:27 |
| Vika (she/her) | Why does specifically Nautilus have its PATH reset only when launching via GIO?
| 17:10:41 |
| Vika (she/her) | This might be a systemd/d-bus thing. | 17:11:07 |
| Vika (she/her) | ok new theory: something resets/doesn't properly export PATH and anything started via D-Bus (and therefore anything launched via GIO) doesn't have its PATH set.
| 17:12:37 |
| Vika (she/her) | aha. dbus-update-activation-environment PATH does the trick. Looks like I forgot to call it in my sway session script.
| 17:18:38 |
| Vika (she/her) | i'm really stupid sometimes. | 17:18:46 |
|  pbsds changed their display name from pbsds (federation borken, may not see reply) to pbsds. | 20:38:54 |
| 20 Nov 2023 |
 piegames | Gnome extensions should be good to go now | 10:00:55 |
 @ThorHop:matrix.org | ElvishJerricco Since you asked in "Off Topic", I decided to try out a few GNOME extensions that replicate the window stacking effects that are found in Windows. Ended up settling for one directly inspired by it. I like this one because like in windows Win+Left and Win+Up will put one window on 1/4th of the screen. You can add other columns and rows, but I like this standard. Hope this helps! -> https://github.com/fmstrat/wintile | 15:26:44 |
|  @iromanyshyn:matrix.org left the room. | 17:29:26 |
 hpfr | because of https://github.com/nix-community/home-manager/issues/1011, I know GNOME Wayland doesn’t source ~/.profile. but I can tell that /etc/profile is sourced at some point because Alt+F2 printenv shows things it sets. I have looked through NixOS modules quite a bit but I can’t seem to find where a bash login shell is spawned under my user. does GDM do this? | 18:58:10 |
| hpfr | * because of https://github.com/nix-community/home-manager/issues/1011, I know GNOME Wayland doesn’t source ~/.profile. but I can tell that /etc/profile is sourced at some point because `Alt+F2 printenv` shows things it sets. I have looked through NixOS modules quite a bit but I can’t seem to find where a bash login shell is spawned under my user. does GDM do this? | 18:58:45 |
| hpfr | * because of https://github.com/nix-community/home-manager/issues/1011, I know GNOME Wayland doesn’t source ~/.profile. but I can tell that /etc/profile is sourced at some point because Alt+F2 printenv shows things it sets. I have looked through NixOS modules quite a bit but I can’t seem to find where a bash login shell is spawned under my user. does GDM do this? | 19:00:10 |
| hpfr | from what I am reading upstream gnome Wayland doesn’t source /etc/profile either, which makes sense, but on my nixos machine it sure seems like something is sourcing it | 19:06:16 |
| hpfr | and since this workaround works it must be run as my user https://github.com/nix-community/home-manager/issues/1011#issuecomment-606869802 | 19:07:01 |
 uep | add a line in there to dump something like pstree to a tmpfile each time it's run? | 21:25:30 |
|  @git_lit_mit_ohne_dir:matrix.org joined the room. | 21:37:29 |
| hpfr | yeah that’s my next move when I get back to it I guess | 21:48:48 |
| 21 Nov 2023 |
 Hubble the Wolverine (they/them) | Heya!
I'm having trouble trying to figure something out
When I try to login via gdm, it keeps logging into gnome X11 by default, despite the bottom right selector (which has GNOME selected instead of 'GNOME (X11)')
I've moved to the gnome branch now but this was still happening while I was on 23.05 | 03:16:51 |
| Hubble the Wolverine (they/them) | This is an issue because I don't have a X11 config set up. When I log in with Nvidia hybrid graphics, it gives me lots of errors and returns me back to gdm | 03:18:01 |
 vcunat | bobby285271 Jan Tojnar the gnome 45 update isn't catching the 23.11 release, right? At least that's my understanding of what the release manager(s) thought yesterday. Branch-off was delayed, but it should happen in about 12 hours - and anyway, final release is in about a week already. | 06:54:07 |
| vcunat | I don't mean to rush you. If it isn't considered sufficiently stable now-ish, that's how it is, and it can get to nixpkgs master later when ready (but not 23.11). | 06:58:58 |
| Jan Tojnar | vcunat: not sure, I guess it is up to release managers. From my end, it is stable enough, the only blocker I saw was the electron build which should now be fixed (currently building on community builder). | 08:30:01 |
| vcunat | In reply to @jtojnar:matrix.org
vcunat: not sure, I guess it is up to release managers. From my end, it is stable enough, the only blocker I saw was the electron build which should now be fixed (currently building on community builder). I think that part is mainly a communication issue. | 08:31:28 |
