The Linux Thread - The Autist's OS of Choice

Is Wayland still shit? I am contemplating switching my neckbeard x11 i3wm setup to Sway when I finally upgrade my PC. But I actually want something functional and that doesn't need a million workarounds to do stuff. I contemplated hyprland but it gives me troon dev vibes, and the stupid eye candy of it seems masturbatory.
Sway is a drop in copy of i3 and is made by the same people that make wlroots, a compositor library used by every standalone Wayland WM except Hyprland and Niri, meaning it should, in theory, experience fewer breaking changes than other compositors. You'll need to fuck around with xdg desktop portals to make shit like screen sharing work. Tbh if you want something fresher than stock X, you might be better off switching to XLibre instead of Wayland.

As far as Wayland compositors are concerned, Hyprland probably has the most normal dev and community around, followed by Niri, then Sway, then smaller ones like Mango or DWL for suckless autism.
 
kind of surprised how many inaccuracies were in this video, and im only halfway done with it
>focusing on the dynamic pricing aspect of net neutrality instead of what it actually was which was isp lobbied red tape (not including the dynamic pricing) to perpetuate a monopoly by keeping smaller isp startups out of the market
>the companies headquarters are in taiwan/vietnam/laos so that means they aren't using rebranded chink hardware with backdoors or other security holes
nationalizing the router market is really the only sensible path forward. you can at least regulate your own country's cybersecurity, you will have extreme difficulty doing that to a foreign country.
whether it's good or not is irrelevant. you have to create a situation where you can only blame yourself rather than a malicious foreign actor. and naturally cves will happen on american made routers thats just the nature of software.
that being said ive got a little lichee pi 4a that can run openwrt so i might toy around with that. do the whole make your own router autism and to tie this back into *nix discussion
im not sucking trumps dick or anything but ive been calling for this for a while after i started seeing chink router cves pop up every few months, and not just from cisco.
i dont like the idea that our home internet infra is potentially being spied on by little proprietary black boxes
 
Is Wayland still shit? I am contemplating switching my neckbeard x11 i3wm setup to Sway when I finally upgrade my PC. But I actually want something functional and that doesn't need a million workarounds to do stuff. I contemplated hyprland but it gives me troon dev vibes, and the stupid eye candy of it seems masturbatory
lmao. when the comparison is sway, or hyprland one of those definitely is troondev, and it's not hyprland.

here's the sway developers blog.




HOLY SHIT lmao. I didn't realize he now has an article mentioning us. lmao.


oh man. I'm sure someone in the open source thread has talked about this one. but I've got to post it there.
 
i dont like the idea that our home internet infra is potentially being spied on by little proprietary black boxes
I'm always surprised how much people don't really care. Hell, look at the Residential Proxy trend in DDoS. Dangle a couple bucks in front of people to backdoor their own networks and you'll have redditors posting about how to min/max using ten different services to get the best payout.
 
they both use libinput for key input. (unless you are going out of your way to use something else on X for key input). as well as touchpad. and mouse input.
The underlying liblary is irevelant, what matters is the interface. And X since there is singular interface provides consistent standard. Wayland does or rather if you do not have any edge cases such as wanting to recieve input key input out of focus(thus OBS problems) and a lot of it is compositor dependent and thus inconsistent. Ofc the manner that X does it is utterly insecure yet the way that wayland niggers did it is even worse as it gimps functionality.
 
The underlying liblary is irevelant, what matters is the interface. And X since there is singular interface provides consistent standard. Wayland does or rather if you do not have any edge cases such as wanting to recieve input key input out of focus(thus OBS problems) and a lot of it is compositor dependent and thus inconsistent. Ofc the manner that X does it is utterly insecure yet the way that wayland niggers did it is even worse as it gimps functionality.
I could be not broken tomorrow if it wasn't for bikeshedding. They could just decide that there is a globalhotkey protocol and that is the way everything can handle hotkeys. At that point if something didn't handle it with the global hotkey protocol they would be the outlier. There isn't anything with the design of Wayland itself that necessarily makes it a problem. It could be pretty easily solved, and could have been a year ago or two or however long people have wanted it. It's not a technical problem is my point. The problem is the people that are the decision makers for Wayland getting in the way of their own project.

Given the choice between everything always having access, or everything being restricted, restricted is the better choice. But if everyone working on the restricted option is too full of themselves to just compromise and agree on how to handle the protocol for allowing access for specified apps that you do want to allow access to everything it then makes the restricted option incomplete.

Though to be at least a little fair to them, if we are just talking general use, my guess is the number of people that use global hotkeys is fairly low compared the number that don't so I can understand why in the beginning there was nothing directly planned for it. I know I've never used them on xorg or wayland. I wouldn't even know they existed if it wasn't for people pointing out that they don't work for obs on Wayland (outside of hyprland and kde) But as it stands with people pointing out it needs to work they should just agree to something and mark this as a solved problem and move onto things that probably need more work.
 
I could be not broken tomorrow if it wasn't for bikeshedding. They could just decide that there is a globalhotkey protocol and that is the way everything can handle hotkeys.
Hopefully Wayland will move through the W3C-WHATWG pipeline (browser developers deciding the standards body is retarded and starting their own) as compositor implementers grow weary of freedesktop.org faggotry.
 
I could be not broken tomorrow if it wasn't for bikeshedding.
That would require free desktop tards to take the task of making X replacement seriously instead of pushing half finished product on everyone. Product that never needed to exist in the first place and one that caused countless man hours to be wasted fixing what was broken by it.
Given the choice between everything always having access, or everything being restricted, restricted is the better choice
Complete nonsense, security is secondary to functionaly as the most secure program is one that doesn't run at all. Security must addapt to functionality not the other way arround. And gimping functionality for nothing more than security theater as UNIX is compromised the moment malicous program is ran by a user is utterly retarded.
 
And gimping functionality for nothing more than security theater as UNIX is compromised the moment malicous program is ran by a user is utterly retarded.
The only legitimate use case the Wayland-tier isolation has for the average user is probably running suspicious proprietary apps. I absolutely do not trust shit like Discord, Teams or Slack to not exploit a builtin keylogger. Said it once and I'll say it again, once xnamespace is functional, Wayland's entire security argument becomes a moot point.
 
Last edited:
But that's just arches version of the same thing other distros do with system configuration files. They pretty much all more or less do the same thing. The only one I think handles it more cleanly is gentoo. Which keeps old versions in /etc/config-archive so you can go back and get older versions before making any changes on updates.
No. Any Debian based distribution explicitly prompts you to choose what you want to do with config files, including looking over a diff before doing so. Because that's the rational thing to to do.
 
i get your point but like what's the better alternative? just lose the security mindset and fix cves as they crop up?
why not start from best practices and have fewer cves in the future?

Complexity does not equal to security nor general reliability, a convoluted methodology fosters idiotic decisions that fuck developers or commercial systems sideways. For instance making your whole system work like a convoluted github repository (Fedora Silverblue, etc.) doesn't make it truly retard proof, it just makes actual maintenance require an internet connection which in of itself sounds like a security nightmare.

Even engineers need a level of human readability, without it they will burn out and make reckless errors.
 
Last edited:
lmao. when the comparison is sway, or hyprland one of those definitely is troondev, and it's not hyprland.

here's the sway developers blog.




HOLY SHIT lmao. I didn't realize he now has an article mentioning us. lmao.


oh man. I'm sure someone in the open source thread has talked about this one. but I've got to post it there.
Oh it's THAT guy. I should really have done a bit more research other than vibes.
You'll need to fuck around with xdg desktop portals to make shit like screen sharing work. Tbh if you want something fresher than stock X, you might be better off switching to XLibre instead of Wayland.
Sounds like I might want to consider XLibre for now then because I cannot be bothered to fuck about with too much stuff. I assume it's a pretty easy drop and replace since it's a fork. I'll be using an Arch based distro and they've got a guide for it. I'd have used the Arch Wiki but it looks like they got butthurt about XLibre.
 
No. Any Debian based distribution explicitly prompts you to choose what you want to do with config files, including looking over a diff before doing so. Because that's the rational thing to to do.
it's what they do is functionally identical to basically every single other distro. All of them have basically the same concept with slight differences. Again, the only one I think has any kind of real improvement on /etc file handling is gentoo, which has the normal handling, and prompt that says you have a file that needs looking at in /etc they have the config-archive dir, which as far as I know is unique to gentoo.

There are tools (or an abstraction layer builtup around other tools really) like etc-keeper which i think was originally made for debian which you can use on any distro to at least get the functionality gentoo has by default (kind of, it basically just adds git tracking to your etc dir, since wont have old versions in config-archive like on gentoo).

but for the most part debian doesn't do anything different from arch as far as I've seen, I guess the prompt is different in the package manger? not a huge difference as far as I'm concerned.
 
Hopefully Wayland will move through the W3C-WHATWG pipeline (browser developers deciding the standards body is retarded and starting their own) as compositor implementers grow weary of freedesktop.org faggotry.
From what I've seen that's kind of what has already started. Particularly for things that get backed at the freedesktop level.

Also river not too long ago decided they would implement what basically amounts to a standard implementation of a wayland compositor, to then allow people to write window managers on top of turning it into what xorg does (from the perspective of a window manager developer). So now using that people can write a wayland window manager using the river backend. It's a nice developement as far as I'm concernd. Someone should have done it sooner. Although it's understandable why no one wanted to put in the work.

Now that it does exist though, I hope dwl drops their backend, so they can actually turn whatever you would call their current project into a proper dwm for wayland. Because as it stands I don't think they managed to fully pull it off. If they're able to offload the work for graphical server (or whatever you want to call it). It would definitely be a lot closer to dwm at that point.
 
The problem is that most compositors are dependent on wlroots as a de-facto standard Wayland library. Apart from Weston or tiny toy compositors like Hevel or shko, everything uses either wlroots, Smithay for Niri and Cosmic, or Aquamarine for Hyprland. I would have less issues with it if it was a definitive standard, not "a" standard. DWL, Mango, River, Sway, all of those are wlroots dependent. What River is doing is great, but the base compositor is still dependent on wlr, so despite making things more modular it adds more depenedencies.

I mention Hevel and shko because they show that you can build a compositor in a suckless style without pulling in wlroots, albeit in a really hacky, janky way. But because the people writing neuswc, the library that these tiny compositors are based on, are pretty much all teenagers, you get shit like this when their interests pivot:

unmaintained​

i've decided to step back with my work on shko.
i grew to dislike zig - i hate looking at it, the hype has died, and with these feelings, i dont want to work on shko.
its a pathetic C wrapper, 90% or even greater of the code consists of c interop -
it makes more sense to rewrite it in C at this point.secondly, programs crash, and the runtime leaves no traceback.
or even, it crashes for no reason at all.

i dont want to be associated with this buggy window manager anymore.

i apologize in advance for my incompetency, especially to those interested and contributors to my work.
for a 7 day period, i will search for a potential maintainer to take over this project. if it fails, shko's repository will be archived,
and in advance will recieve no more updates.
open a issue if you would like to maintain shko
For a real suckless wlroots alternative, there needs to be a lot of effort from a lot of people that probably would rather just use wlroots and focus on maintaining their compositors rather than backend libraries. The problem with wlroots, like all Wayland things, is they introduce retarded breaking changes whenever they feel like it because "move fast break things", which has the trickle down effect of making every compositor that depends on wlroots have to sacrifice time and effort to account for upstream breakage. IIRC that was the reason DWL was originally dropped before being picked back up by its current maintainer(s?).
 
Back
Top Bottom