| 12 Sep 2025 |
Sandro 🐧 | I tried enabling fwupd on a headless system and it has many GUI dependencies and is pretty big because of that | 12:08:49 |
Sandro 🐧 | also in my experience only dbx gets updates on almost all of my devices. I had like 3 other updates in 5 years. So support is not that well. | 12:09:53 |
Sandro 🐧 | but I would be fine to enable it by default in the plasma 6 module | 12:12:21 |
Grimmauld (any/all) | it pulls pango/harfbuzz, but that is not what makes the closure large.
The problem is the fact it has a runtime dep on GCC via valgrind (i think) | 12:29:31 |
Grimmauld (any/all) | i mean i guess modemmanager is somewhat large, but at least on systems with NM that already exists | 12:31:32 |
aloisw | In reply to @emilazy:matrix.org pretty not good to not have a way to get firmware updated You still have to trigger the update manually. It doesn't seem like installation of fwupd is the weakest link here. | 15:23:03 |
Lun | idk if this matters but every time it runs it makes amd dgpus do a fan spin up spindown cycle | 15:24:05 |
Lun | * idk if this matters but every time it runs it makes amd dgpus that were otherwise suspended do a fan spin up spindown cycle | 15:24:26 |
aloisw | "it" being the update checking command? I guess it needs to power on the card to read its firmware version. | 15:26:42 |
Lun | The scheduled check for updates one (fwupd-refresh?) iirc. It's not really avoidable but it might be an annoyance if it was defaulted on for some laptop users | 15:28:53 |
Lun | well it is avoidable but it's not fwupd's fault or avoidable on its side, amdgpu needs to be less silly about this and there's been a fix a few times on the list and it's never been applied so this'll presumably stay broken forever | 15:29:49 |
Sandro 🐧 | I think it would require to create a libnm to make it smaller | 15:43:39 |
Sandro 🐧 | I only know RedHat which has NM on headless systems... | 15:43:49 |
K900 | Splitting libnm would be nice actually | 15:44:31 |
K900 | Because Jovian needs 32 bit libnm | 15:44:35 |
Sandro 🐧 | https://github.com/NixOS/nixpkgs/pull/241762 ? | 15:48:29 |
Sandro 🐧 | * I started something small a while ago in https://github.com/NixOS/nixpkgs/pull/241762 | 15:48:36 |
| 13 Sep 2025 |
@aidalgol:tchncs.de | I have a python GTK application that is not dark themed. It appears to be using GTK 4.
(process:2118965): Gtk-WARNING **: 16:52:06.282: Unknown key gtk-modules in /home/aidan/.config/gtk-4.0/settings.ini
| 04:58:21 |
@aidalgol:tchncs.de |  Download image.png | 04:59:00 |
@aidalgol:tchncs.de | My Plasma GTK settings seem... wrong. | 04:59:15 |
@aidalgol:tchncs.de | If I click the dropdown menu, there is a single option "Breeze". | 04:59:38 |
@aidalgol:tchncs.de | Do I need to add a dark GTK theme package to my system config? | 05:00:08 |
aloisw | Have there been Qt-related changes in the last staging-next? Nheko now doesn't show replies properly any more and the context menus open in the top left corner. | 08:02:38 |
aloisw | (Not really KDE-related, but I figure the folks knowing about Qt are probably here.) | 08:03:08 |
K900 | 6.9.2 update | 08:15:36 |
K900 | Not sure how that would break nheko though | 08:15:44 |
aloisw | They seem to do cursed shit, it has broken on Qt minor updates before. | 08:19:19 |
K900 | Consider: neochat | 08:19:43 |
aloisw | Is it better now? Last time it couldn't load more than one or two screenfuls of backlog, even when more was unread. | 08:21:28 |
aloisw | (Also I guess I could just use Element Web if I need to get comfortable with a new client anyway, at least that doesn't depend on libolm.) | 08:22:16 |