Out of curiosity, why not just leave ssh access to the local network so you can only reach it by VPN in the first place? Note I might be misunderstanding what the goal of this was, so feel free to lmk if I’m off the field with my question lol
I was surprised to learn this was a thing, impressive, however;
‘the VPN app sends a request to the VPN server to open a random port’
‘the active port number will change when you disconnect and reconnect the VPN.’
This will not work OOTB with Plex for example, you would need to change the port in the app every time. It becomes difficult to serve anything statically, like a XMPP server or anything that doesn’t let you configure the port.
You also would need to be at home to check which port you’ve been assigned eg if the connection drops and you get assigned a new port, defeating the whole ‘remote access’ thing.
What price bracket are you looking at? The two laptops that I normally use in that situation is a used Thinkpad X1 Carbon I got on eBay, and a HP Dev One that works pretty well for that.
The Thinkpad link that was shared below looks pretty nice, they tend to be fairly cheap and easy to get replacement batteries and parts. There’s a lot available in that $150 to $200 bracket on eBay. Edit: I just saw it’s 14", so a bit bigger than what you wanted. You can filter by screen size and price on eBay to give you an idea of what you can get. You may need a new battery depending on the age, so keep that in mind.
I hate moving windows around.
All windows open maximized without window decorations. Meta+WSAD moves the active window to the upper/lower/left/right half of the screen.
Meta+PgDown minimizes, Meta+PgUp maximizes. Meta+Q tiles windows horizontally, Meta+E vertically. Meta+X closes the window, Meta+Spacebar shows the desktop, Meta alone shows the workspace overview.
Fuck hunting for window borders, clicking and dragging. And fuck configuring all this in a text file.
No. I need the functionality of a full desktop environment.
And KDE’s workspace overview is awesome. One keypress and I see all open windows, all workspaces and a global search field that switches to a program when it already has an open window and opens a new window if not.
And a tiling WM on top of KDE would be pointless to me since the behavior of a tiling WM can be configured through the GUI in KDE without installing anything extra.
I’m not familiar with sway or hyprland, but KDE automatically finds and configures any modern scanner and printer in the network, makes all programs use the same theme, saves my passwords and certificates, auto-mounts attached drives, auto-starts programs and services, handles which program opens which file type, has a nice workspace overview, lets me configure the firewall, grub, bootsplash screen, VPN, network settings, monitors, keyboard layout, etc… all with sane defaults out of the box, localized to my language, and easy GUI configuration.
For what it’s worth, a large number of the things you listed are actually portable into Sway, i3wm, and a lot of other tiling wms just by way of running the KDE settings daemons - I do the same kinds of things (network printer, theming, auto-mount, auto-start, XDG config, firewall, vpn, network settings, monitors, keyboard layout) just by having i3wm start up xfce-settings-daemon.
I’m not familiar enough with KDE to make promises about grub and splash, but I would imagine those would also work exactly the same as well. In fact, a little bit of searching and it looks like if you’re on Wayland you could even just replace KWin (the KDE window manager) with Sway in the startup files and be 95% of the way there. Might just need to configure a system bar or something to that effect.
Yeah I also use KDE on my desktop, though I have my laptop running QTile because the tile hotkeys are much more convinient than navigating with the trackpad.
Sure, but that’s not the only benefit to having full control over the entire tiling interface. I enjoy building out the features and visuals I want in python. It’s fun to have that level of control.
GNOME Shell 45 moved to ESM (ECMAScript modules). That means you MUST use the standard import declaration instead of relying on the previous imports.* approach.
<span style="color:#323232;">import Clutter from 'gi://Clutter';
</span><span style="color:#323232;">import Gio from 'gi://Gio';
</span><span style="color:#323232;">import * as Main from 'resource:///org/gnome/shell/ui/main.js'
</span><span style="color:#323232;">import * as Volume from 'resource:///org/gnome/shell/ui/status/volume.js';
</span>
well that’s what I tried, but how would I know where Volume is located in the path 'gi:// … '. Is there any way of browsing / exploring the jave files to that I can actually know what the path is?
I edited my comment with an example for your code and my best advice for figuring out the path of gnome shell imports is by browsing /usr/share/gnome-shell/js/, the docs are not very helpful.
It is very likely the wrong path, I just extrapolated the path from the gnome-shell git repo. I don’t use Gnome myself, I’m on the enemy team using LXDE on Devuan ;)
Just to clarify. The gi:// resources are GObject Introspection modules which are used for multilanguage bindings to native libraries. On my system, GI modules are found in /usr/share/gir-1.0/ . They’re just imported by name and sometimes version using gi:// (there are examples in the link in my first comment).
As I don’t have Gnome installed I can’t be sure of the path to gnome shell modules imported using resource://, but it’s probably the path I wrote, but without js/.
People use computers to accplish tasks. That requires running software on an OS, but nobody runs software or an OS just to sit & watch it exist. They run it to accomplish tasks.
Different distros mostly vary in how easy it is to accomplish various tasks. No one distro is the easiest for everything, so people make different choices depending on their needs.
Aha, I think I misunderstood your situation then? I assumed you’re running these routing rules on your client machine, so you’re able to access your ssh server without it going over the VPN – not that your server is running a VPN active that blocks external connections…?
But if I didn’t misunderstand, I’d mean the (assumingly static) ssh server’s IP.
It has a structured yaml with a test command for potentially destructive config changes over ssh. Other than that: none. It was a real pain upgrading some servers, as always with Ubuntu.
It literally has no benefits, and is only a pain to use.
Actually, it does have one benefit: it integrates with Canonical’s other tech. For example, MAAS uses ot for networking, and I bet lxc uses it somehow.
Can’t even being to agree enough on this. Unless you specifically need something that an SBC - ARM or X86 - offers, a second hand thin client or USFF computer will be a better fit, plus they come with high-quality power supplies and solid cases.
Well, not really. The HP g3 mini is roughly the size of a paperback book and costs around 100€, depending on the specs. Similar devices of slightly older makes are even cheaper.
So, yes, they are physically larger, but still pretty small. Chances are, you don’t actually need a tiny device like a Pi, so you should at least consider SFF PCs.
How do you handle which GPU is used in which game? I would guess you have an AMD iGPU, and a Nvidia GPU for games, right? Maybe something along those lines got updated?
That’s correct, but i mostly let the laptop handle it, unless i know for a fact that a game needs/doesn’t need the N GPU, in which case i either manually switch it over or (and this is the case for wine apps through bottles) i configure the program to only use the iGPU
What is even the value of Netplan on… desktop? Most people just pick their WiFi in the menu in Gnome. That sounds like a lot of unnecessary complexity.
For servers, sure, it’s fairly nice. But, desktop? Why?
For Ubuntu 24.04 LTS, Canonical plans to polish the Netplan codebase and deliver a Netplan 1.0 release with API/ABI stability. They are also hoping other Linux distributions begin adopting Netplan. Debian so far has decided to go with Netplan for their nework stack on Debian Cloud images.
That’s probably the reason for pushing it to desktop builds.
If you’re just using DHCP, you won’t. What Netplan does is take a YAML input file and renders it as a systemd-networkd or NetworkManager configuration file. It’s a very quick and easy way to configure your network, and even have a try command that auto reverts in case you get kicked out of your SSH session.
It seems like what they’re doing for the desktop is hacking up NetworkManager so that it saves back its config as Netplan configs instead of regular NetworkManager configs. That’s the part I’m confused about, because NetworkManager is huge and Netplan doesn’t support close to every option. Their featuresets are wildly different. And last time I checked, the NetworkManager renderer was the least polished one, with the systemd-networkd one being close to a 1:1 match and more reliable.
It made a lot more sense when it was one way only. Two way sounds like an absolute mess.
linux
Active
This magazine is from a federated server and may be incomplete. Browse more on the original instance.