Max-P
A bit of an extra tidbit of information, that’s why things like /usr/local exists: to keep things separate from the distro provided packages. A lot of things have directories in /etc for the sole purpose of isolating “what the distro provides” vs “what the user provides”. Often you can mask distro-provided settings by symlinking the same in /etc to /dev/null.
So it’s best to assume /usr is managed by your package manager and you shouldn’t touch it because it’ll be overwritten on next update. Or in the case of immutable distros, you can’t even write there yourself.
Plus when you do backups and restores, you can backup /etc, reinstall your distro, then restore your /etc and you should mostly end up back where you were.
Don’t overwrite system files managed by pacman and it won’t break your setup.
Put your script in the corresponding place in /etc where it’ll be handled properly, or better yet, do it correctly and set it up directly in your xorg.conf config. Or just use localectl set-keymap dvorak which will do the same thing but also handle non-X11 use cases such as the virtual consoles and Wayland.
KWallet most likely
It’s a Threadripper system which effectively behaves like two CPUs and loads of coree, two GPUs, one dedicated to my desk for the monitors, and the other one can be reassigned freely with VFIO to be a few different things. The TV is connected to that GPU. Storage is all ZFS.
- One VM is a kube node to run stuff on that GPU
- One VM is the media center / gaming stuff
- Technically I have a Windows and Mac VM too but I practically never use them.
When the second GPU isn’t attached to a VM, I can also use it on the host with DRI_PRIME. The host is also a kube node, so I can also run some (modest) AI stuff there too.
The rest is random glue scripts like detecting when the controller connects and shuffling VMs around on that signal. The kube stuff is brand new, half the things are just regular docker compose files still.
I’m looking into trying out kubevirt and see where that goes. The GUI is the only thing left that’s relatively normal on the host and I’d very much like to make that a container and split things up in sort of “activities” so the browser is its own thing, each project is its own thing so I don’t npm install a rat.
I’d stay clear from those “clients”. I had a quick look and holy shit, they have stores with ridiculously predatory prices for skins. That’s 100% exploiting kids and their parents. Fuck that. Nope nope nope.
Use Prism. It’s open source, has existed for a very long time, works well and has been the goto on Linux for a while, and is actually a real launcher.
Generally, those people are experienced users that know exactly what they want out of a distro and don’t really need help for anything. Those distros usually do a few things that the user is seeking.
For example, for some people, typing their thesis in LaTeX using emacs is the better workflow. To any average person that sounds insane when Microsoft Word is so easy to use and does the job just fine. But they enjoy it, it works for them, paper gets written, everyone is happy.
Distributions are a spectrum between novice users and expert users. Some people want to put the USB in and be good to go. Some people want a very precise setup for very specific needs.
You may ask, why not start with Ubuntu/Mint/Pop and remove what you don’t like? Well, it’s much easier to start with a blank slate than making one by chopping everything out. For my particular use case, I moved to Arch in big part because I got tired of the mainstream distros getting in my way, and wanted to start the other way around and only install and configure what I want, the way I want it. So Arch for me.
I know experienced users that really don’t care about messing around and are happy with how it runs out of the box and are happy with the development environment provided by something like Ubuntu/Fedora.
And then there’s my box which is a NAS, a workstation, a media PC for the TV, a build server, and a few other things, and it’s all dynamically reassignable. Friend can pick up the controller in the TV room and a GPU gets assigned to it and starts up Steam in Deck mode on the TV, while I can still do my stuff and game on the workstation side for local multiplayer. If the game needs a server, no worries, it’s a kube node, I can temporarily transfer the server locally and back on one of my real servers. Guest needs a PC? Sure, take this monitor and this keyboard, here’s an ephemeral Windows install. Sure, I could probably twist Ubuntu into doing all that, but it’s one hell of a lot easier starting from scratch.
Both GTK and Qt have good Python bindings.
GTK example: https://github.com/Taiko2k/GTK4PythonTutorial
There’s also PyQt but it looks more complicated and I couldn’t find as nice and straightforward of an example as GTK but I found this: https://realpython.com/qt-designer-python/
If you want to go to C, GTK works about the same way. If you want C++, Qt is pretty good there.
Otherwise you can go SDL and just put whatever pixels you want on the screen on your own.
DeepFilterNet is fairly effective. Not quite as good as RTX Voice I would imagine, but it does wonders for my meetings with the duct right behind the webcam and the furnace right behind that wall.
EasyEffects supports it out of the box as long as the plugin is installed, and the Flatpak also comes with it bundled I believe.
That said, for recording, I would recommend applying the filter in post so you have the original to try different settings and get the best quality.
