I’m gonna comment and say that’s the point.

You start out with bare minimum and install what you need. As you go you generally have an idea of what is and isn’t on your system. It’s not as annoying as Gentoo with all source compiling, not as anal as nix.

If something breaks, you go to ArchLinux.org and 95% of the time it’s mentioned on the front page so you follow the instructions and move on. It’s a very transparent distro, little drama to follow unlike Ubuntu/canonical or fedora/redhat.

It used to be harder to install and which gave some street cred, but they simplified it a bit which is nice.

The Stans give an unbalanced look at arch. I use arch because I want the latest packages, I don’t want to segment my packages between my repos and tarballs when there’s a game stopping missing feature on a package pinned to a 2yo version. I don’t want to learn a whole scripting language to carefully craft my OS like nix either. I want a current OS that’s easy to fix and easy to install packages so I can go back to what I was doing.

I had only used kde once before like 7 years ago and I wasn’t a huge fan. I wanted to try it again and I honestly really like it over gnome. I usually go tiling but felt lazy with a new laptop. The trackpad gestures are really solid.

nmcli is quite nice actually. My only real issue with NM is keeping track of what it’s doing behind the scenes.

So I want and have ip forwarding, and I only want to make a firewall whitelist between two of the interfaces.

I’ve uninstalled iptables, nftables isn’t running, NM has the firewall backend disabled, and ip forwarding is on.

This should result in traffic moving between the interfaces, yet traffic is moving between two of the interfaces, and blocked between two of the interfaces. It just doesn’t make sense.

Thanks for the suggestion, but I’m using NM for managing the AP and managed connections, not so much the bare connecting to wifi things.

The only real alternative to NM in this situation is a handful of delicate config files for iwconfig and dnsmasq.

A lot of software wont be distributed with a PPA to add.

Additionally, debs are useful for offline installations, with apt you’re able to recursively download a package and all of it’s dependencies as deb files, then transfer those over to the offline machine and install in bulk.

That being said I’ve never had great luck with the software center, it’s always felt broken. I’ll typically just dpkg -I .

I think I first installed linux some time around 2009. I’m only just now starting to contribute to libraries, unrelated to linux. Its such a cool feeling growing along side the open source movement.

Polonium

Hm I’m not sure if that’d really give me what I’m looking for. I know its certainly possible to configure KDE and Polonium to get me 90% there but I think I’d rather just have a normal floating setup I can switch to if need be. I’d need to remap a significant amount of keyboard shortcuts that would stop making sense in the context of a full floating DE.

I really just want a very fast app launcher like dmenu, dynamic tiling, and monitor independent workspaces. I have a particular setup using certain alpha keys for my workspace.

I never really enjoyed the experience of tacking things onto an existing DE and having to mess with UI configuration. I’ve been really loving XMonad for a few setups and my ideal wm would be something that’s extremely low power and low fluff. Even if I only eek out 10% more battery life, breaking the 10hr mark is more valuable to me than most bells and whistles.

I’m just really lazy. I could load up my xmonad setup in 20 minutes but I wanted to see the state of wayland and that requires learning a new wm’s configuration quirks.

I’ve been using gnome as a “base” DE for years, what that means is I install it, then install my tiling wm and use all the gnome utilities.

I recently had to set up a few new machines and decided to try KDE on a couple and I’m really enjoying it. I haven’t even gotten around to installing a tiling wm because I want to learn a wayland option and that’ll take some time. I haven’t ran into pain points listed here but one thing I like is when I want to do X, there’s usually already something ready to do X for me. Years of gnome and I felt like the devs were always fighting me. I haven’t really used a full gnome setup in a few years though, but I know the “mommy knows best” attitude is still prevalent with the devs.

I was using it on a new work machine, it was fine.

The main issue is all the good tiling wms are X11 based and I don’t really want to use a wayland version of i3. I want some dynamic tiling goodies.

postin’ from my 4th gen X1 Carbon running arch converted from antergos.

So what the ram is soldered, its 4lbs and still gets 7+ hours of battery life after 8 years of use.

Yeah I actually just prefer the command line, I’ve never had to force myself to use it. I even tried using VSC for a bit recently but i couldn’t get myself to like it. I just use nvim with some plugins in a tmux session now and its productive as hell.

Of course I don’t browse the web with the command line. For merging branches, I always merge main into the working branch first, check conflict files, and go through the file finding the diffs and resolving them. I’ve used merge tools before that were sorta nice but I had my own issues with them.

Maybe it’s the type of programming I do. I don’t do any web stuff, so file count is down. For larger code bases I keep a non editor terminal up and will grep -re for word/phrase searching, find to look for specific files, etc. I’ll occasionally use an IDE, typically eclipse based because embedded, but I don’t find myself missing the features they add.

Thanks for the explanation, that does sound useful.

That’s fair, there’s plenty of uses for source control.

I was speaking from a programming context though, as this is a programming community.

I really never understood why one would need a GUI for git except for visualizing branches.

I feel like I’m crazy seeing so many people using clicky buttons for tracking files. I need like 4 commands for 95% of what I do and the rest you look up.

You’re already programming! Just learn the tool!

And now there’s a github CLI tool? I hate to beat a dead horse but Microsoft pushing their extended version of an open source tool/protocol is literally the second step of their mantra.

A typo in the first paragraph of the article in a wiki wont make the 5th paragraph tear down the entire wiki.

I should take another look into it. Thanks!

Web assembly isn’t quite the same as a js frontend though, is it?

It’s typically for complex single page apps and has some weirdness with normal usecases, no?

I could be wrong but I was looking into it a few months ago and it seemed immature.

a front-end language

I love rust but this requires killing the web app and using basic html. which i’m also pro.

The endless packaging solutions for python is exactly the flaw that they’re talking about.

Packaging a python application to work over an air-gap is extremely painful. Half the time its easier to make a docker container or VM just to avoid the endless version mismatch pain.