The Linux Thread - The Autist's OS of Choice

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
Note the first point. Some compositors don't have that concept, but that means that some compositors will. Because of practical considerations. Now you have an off-standard feature that users want, but it's entirely dependent on how the authors of the compositors implemented it. Have fun creating programs for this kind of schizo environment.
This is the fundamental reason Wayland exists and the fundamental problem with it at the same time. Wayland is not trying to address the same problems as X11. It operates on a lower, more abstract level. This allows it to serve a wider variety of platforms and usecases, but also means that it doesn't natively include alot of functionality needed in common situations. Realistically it should've shipped with a partner project explicitly extending Wayland for traditional desktops, potentially with similar projects for other platforms like terminal servers, mobile, etc. It is a much stronger foundation but their refusal to properly support the actual userspace, paired with stupid internal politics led to an extreme delay of adoption. I'd imagine one of the reasons that wayland is actually seeing use now is because libraries like wlroots actually built a usable bridge needed for usespace implementations.
 
I'd imagine one of the reasons that wayland is actually seeing use now is because libraries like wlroots actually built a usable bridge needed for usespace implementations.
My problem with Wayland isn't even that I have some esoteric use case where it somehow has completely broken major things I want to do. It's literally the social bullshit and the attitude of the devs that makes me expect it will blow up in some major way and if it does, I'll feel like a goof for adopting it.

I like Xfce (although they are making noises about adopting Wayland). As far as I'm concerned, the less there is to a DE the better.
 
I like Xfce (although they are making noises about adopting Wayland). As far as I'm concerned, the less there is to a DE the better.
Xfce Wayland is very much under way.
Every single Xfce application runs under Wayland just fine (thanks to GTK), the only thing not done is porting xfwm4 to Wayland (which is going to take a few years according to the devs).
For now you can use labwc as a Wayland compositor and it works! Its just alpha quality and major things like desktop scaling are broken on Wayland because its just come out. But it is being actively worked on at least.

If you want a more basic desktop running under Wayland, try LXQt, its past the alpha stage that Xfce and Mate seem to be stuck in.
LXQt does something neat with compositors too, you can choose several (off the top of my head, they had labwc, Wayfire, Hyprland, Sway and Kwin, maybe more). I remember the Kwin session working quite well, but KDE is still the best if you need Wayland for whatever reason.
 
Surfing the web to see what old unix fags say. jwz's blog:
  • Wayland is not a window system, it is a collection of hundreds or thousands of "extensions" that any given "compositor" (window manager) may or may not implement, à la carte, meaning there are actually dozens or hundreds of "Waylands", and it's amazing that anything works at all."
 
(On a discussion as to why Wayland has no mechanism to define a primary monitor)
Actually fucking delusional.
That setting is specific to KDE's environment. I don't think non-KDE apps have a standard way to figure out which monitor is the primary (for things like status bars and docks).

Note the first point. Some compositors don't have that concept, but that means that some compositors will. Because of practical considerations. Now you have an off-standard feature that users want, but it's entirely dependent on how the authors of the compositors implemented it. Have fun creating programs for this kind of schizo environment.
This is the biggest fucking issue. People can bitch about the underlying structure of X11 all day long, but from the application development perspective, you had a solid separation of the display server, the window manager and the actual applications. You could run a single application that took up the whole screen without a window manager.

Combining the window manager and composer was like the fucking Itanium. Let's just build scheduling into the compiler, that way when we come up with better scheduling algorithms, we just update the compiler! Yea, great. Except both of those things are insanely complex and difficult. CPU architects write scheduling in their own tools which are totally different than what compiler authors are use to and visa-versa. Now the compiler author has to think, "Okay, I told instruction batch 1 to divide, now I can't send it another division command for five cycles and that's when I can read the result." EPIC wasn't even a pure implementation of VLIW, and it was still retarded.

Not having a standard fucking way to do something across all environments is insane. Tiling window managers in X11 can use an x window property to know if some windows should be dialogs. Wayland has no such attribute at all. This meant on Sway, the file transfer dialog in Dolphin file manager was treated like a full window instead of floating. You needed custom regex rules on the title just to make sure the file transfer dialog didn't split your windows. This isn't an issue with Hyprland and I'm not quite sure how they worked around it.

Wayland reduces the overall number of possibilities. Building new managers was a task in and of itself. Now you have to build it with a composer?! Why not just have it do one thing and not have to debug endless shit in wlroots, aquamarine or whatever the fuck you're using? It's core design makes it more difficult to write complex applications that work across every different composer. Trivial applications like chat and Gimp are fine, but clipboard managers, screenshot tools, recorders, workspace managers, KiCad .. they all have to support each individual environment or portal APIs that may or may not exist (XMPP component vibes here).

It was a shit decision and the core Wayland devs have never acknowledged what a fucking terrible idea this was. I hope XLibre gets HDR support, patches mpv for Xlibre+BT.2020 and years from now we're seeing Gtk and Qt reattach XLibre support because it works better than the Wayland dumpster fire.
 
That setting is specific to KDE's environment. I don't think non-KDE apps have a standard way to figure out which monitor is the primary (for things like status bars and docks).
xrandr | grep "primary" will get your primary display info from X11, regardless of your DE. But apparently that's old and fascist: random monitors are the new hotness.
 
You are also not limited by n00b-friendly Linux distros in most respects. Here you can see the OS statistics of every single top 500 non-distributed supercomputer in the world
Yes, you aren't limited by "n00b-friendly" distros when you completely ignore all the stuff that supposedly makes them friendly and just run a server install as a basis for your highly specialized computational software, because at that point most of those distros just turn into the equivalent of a Debian netinst image.

Surfing the web to see what old unix fags say.
Two funny recent posts by OpenBsd developer Ted Unangst:
 
Last edited:
I've seen a couple of videos referencing this, and I'm tempted to print it out and frame it.
View attachment 7582573
(On a discussion as to why Wayland has no mechanism to define a primary monitor)

It really does sum the attitude up nicely.
that guy in particular is a retard. when you see people saying dumb shit like this in wayland discussions. look at the the names.

almost every time. it will be this dude saying it.

honestly wayland would be so much further if they removed this retard, instead of kicking out people that try actually implementing things. like vaxry from hyprland.
 
This is the fundamental reason Wayland exists and the fundamental problem with it at the same time. Wayland is not trying to address the same problems as X11. It operates on a lower, more abstract level. This allows it to serve a wider variety of platforms and usecases, but also means that it doesn't natively include alot of functionality needed in common situations. Realistically it should've shipped with a partner project explicitly extending Wayland for traditional desktops, potentially with similar projects for other platforms like terminal servers, mobile, etc. It is a much stronger foundation but their refusal to properly support the actual userspace, paired with stupid internal politics led to an extreme delay of adoption. I'd imagine one of the reasons that wayland is actually seeing use now is because libraries like wlroots actually built a usable bridge needed for usespace implementations.
that's basically what wlroots is. in a sense.


@HootersMcBoobies
Not having a standard fucking way to do something across all environments is insane. Tiling window managers in X11 can use an x window property to know if some windows should be dialogs. Wayland has no such attribute at all. This meant on Sway, the file transfer dialog in Dolphin file manager was treated like a full window instead of floating. You needed custom regex rules on the title just to make sure the file transfer dialog didn't split your windows. This isn't an issue with Hyprland and I'm not quite sure how they worked around it.
That isn't a wayland specific issue you are talking about the same thing happens on xorg, on i3. The thing sway is replicating. And a lot of tiling window managers. It has nothing to do with the display server or the compositor.

and wlroots is the answer to what you said at the end.
 
Last edited:
That isn't a wayland specific issue you are talking about the same thing happens on xorg, on i3. The thing sway is replicating. And a lot of tiling window managers. It has nothing to do with the display server or the compositor.
You know, it's just occurred to me that since all the actual work is done by the likes of sway & hyprland, all wayland actually have to do is say "Have a ~/wayland.conf file. And in that file have a line that says Primary = HDMI-1-3 2560x1440 590mmx334mm 75hz 8bit_color or whatever connection and specs you need" and boom, the problem is solved. It's then up to implementations like sway to code that if they want to, and applications like factorio to use that if they want to, or just have users type the damn thing in with vi, but you now have a working protocol for defining the primary fucking monitor.
It's neither perfect nor elegant, but I'm pretty sure I just added more functionality to wayland than that entire year long bitch-fest.

I hate these people so much.

honestly wayland would be so much further if they removed this retard
With fire, preferably.
 
Last edited:
This is the biggest fucking issue. People can bitch about the underlying structure of X11 all day long, but from the application development perspective, you had a solid separation of the display server, the window manager and the actual applications.
Imagine that. Imagine a protocol working for *40 plus fucking years*. Then a bunch of retards think forcing some dumb bullshit on you is just going to be tolerated.
 
that's basically what wlroots is. in a sense.
The Wayland guys should have really written something like wlroots from the start.
If they did, Wayland would probably have been in use for 20 years and X would be long dead. But they didn't.
Now, its too late, we now have:
-Gnome Wayland
-KDE Wayland
-Hyprland Wayland
- wlroots Wayland (which everything else seems to be basing off of)
Its almost like Wayland was built to cause the already small Linux user base to fracture even more.
 
1751462900275.webp

 
Oh, the shit icing on the turd cake is that he was arguing with factorio's Linux dev.
:story:
The whole thing (so far) is here, but be warned: Here Be Retards.
Bless the Factorio dev's heart for trying.
I've looked at what it would take to natively support Linux and quickly abandoned the idea. If it works with wine, great; if it doesn't, well.. maybe, when the year of the Linux desktop finally arrives. 🌈
 
Bless the Factorio dev's heart for trying.
I've looked at what it would take to natively support Linux and quickly abandoned the idea. If it works with wine, great; if it doesn't, well.. maybe, when the year of the Linux desktop finally arrives. 🌈
Sorry chud, we care about security. You can't have usability here. I'm off to take my HRT subsidized by Redhat.
 
Bless the Factorio dev's heart for trying.
I've looked at what it would take to natively support Linux and quickly abandoned the idea. If it works with wine, great; if it doesn't, well.. maybe, when the year of the Linux desktop finally arrives. 🌈
I think it might be better to develop a game container platform that developers used, that way the games will work on any device that supports that container. Like if proton was ported to macOS and Windows
 
that's basically what wlroots is. in a sense.
Yep, there should have been a blessed wlroots equivalent from the start. Its no coincidence that we saw a huge influx of adoption to wayland and a number of compositors being developed once wlroots was created and became a general standard for implementation. If the project was properly managed there would have been a sister project to bring it up to feature parity for desktops early in development. There also could have been separate projects for alternative platforms like mobile, kiosks, network clients, and VR. Wayland spent far too long in an abstract bubble hoping everyone else would figure out how to actually develop user space. We are finally in a state where that problem has been resolved, 15 years later than it should've happened.
 
Yep, there should have been a blessed wlroots equivalent from the start. Its no coincidence that we saw a huge influx of adoption to wayland and a number of compositors being developed once wlroots was created and became a general standard for implementation. If the project was properly managed there would have been a sister project to bring it up to feature parity for desktops early in development. There also could have been separate projects for alternative platforms like mobile, kiosks, network clients, and VR. Wayland spent far too long in an abstract bubble hoping everyone else would figure out how to actually develop user space. We are finally in a state where that problem has been resolved, 15 years later than it should've happened.
I think I've said it before but X11 was designed by Engineers who did all the leg work to make everything that was needed while Wayland was designed by marketing and they only had an abstract concept and expected others to put the work into making it a reality.

At this rate Xlibre will leapfrog Wayland within months and make Wayland irrelevant again.
 
Back