The Linux Thread - The Autist's OS of Choice

An ideal solution in my opinion would be some kind of generic server for all forms of input, audio, and video that can be piped wherever it needs to be, usually into speakers, out of peripherals, and onto screens. That way you could set up whatever architecture you need, at the cost of lots of implementation complexity.
It might be an interesting way to handle mini screens that are only good for widgets and such.
 
  • Agree
Reactions: Creative Username
Honestly if you want X11 to be more popular you should try to work improving its benefits. Like, X11 allows the display to be on a different computer then the application. If you could expand on that you could have setups like having a massive supercomputer running intensive apps that get used on basic computers. You could run Blender or maybe even games on your more powerful server then use thos apps from any device with the X11 windowing software installed.

And Wayland does have dramatic improvements, but it's clear it has severe gaps that need to be addressed.
ssh -Y user@server firefox to try it out easily. It works, but performance is a lot worse than just about any Remote Desktop solution because X protocol has a lot unnecessary back and forth.
 
  • Disagree
Reactions: teriyakiburns
The problem with plain X forwarding is that many programs need to be transferred pixel by pixel, which is not fast, especially at modern resolutions. You can somewhat improve this by using the -C (compress) flag with ssh, but it will still not be fast. There are also programs using tricks like being protocol aware and damaged areas. xpra is a program under which you can run other programs, akin to screen or tmux, so you can set a server up that runs specific graphical applications and then log in and out from whatever machine without ever closing the program on the server. Contrary to something like VNC you also don't need to stream the entire desktop, but only that one program so it's basically indistinguishable from a program running locally. It is actually a pretty old program but has seen a lot of updates recently. It now supports hardware video encoding on the host machine which is the most efficient way these days to stream the screen contents of a program over the network and what usual "game streaming services" and such use. I have my own setup like this, using ffmpeg. I play games on my laptop running on my desktop this way. My laptop is an ancient Kaby Lake potato, but it doesn't really matter as long as there is hardware video decoding pretty much any hardware is "good enough" as client. If you can watch videos on it at the screens native resolution without stuttering, it is fast enough for streaming. At least on a local network this is fast enough to even play shooters.

With X there's also nothing inherently stopping you from running multiple X servers on the same machine, on the same graphics card. (driver permitting) I could for example hook up two keyboards, two mice and two screens to my PC, then have one user run ratpoison and the other user KDE and they'd never even know that the other person exist. Since X servers communicating on the same machine is basically memory copies and very fast, you could easily share programs between the two users or even do things like set the mouse cursor up to seamlessly leave the screen of Desktop A and go visit Desktop B, with a program like x2x.

Of course in X there's no inherent requirement for any window manager or compositor, so you can for example also start an X server to only run one program, like a game (makes sometimes sense with windows games ran via WINE who don't behave correctly with some window managers) or a widget for the second screen.
 
In Wayland lunacy news, pcsx2 devs disabled native wayland by default, because they had enough of user complaints over lack of ability to set relative window position, among other things.
https://github.com/PCSX2/pcsx2/pull/10179 Go read wayland protocol merge requests if you want to lose faith in humanity. Years of arguing, which can end with gnome devs vetoing the proposal because of muh security, or simply refusing to implement any of it.
oro1.png

My favorite comment from that thread. A maintainer responds to this snide bitching with "I'll(sic) look forward to your pull request". That would be useful of them, but then I looked at the Github profile.

oro2.png

Those Wayland problems sound like you problems, PS2 emulator developer.
t. a literal hatchling Rust fag (currently learning!), documentation writer and program packager :story:
 
View attachment 5525223

My favorite comment from that thread. A maintainer responds to this snide bitching with "I'll(sic) look forward to your pull request". That would be useful of them, but then I looked at the Github profile.

View attachment 5525224

Those Wayland problems sound like you problems, PS2 emulator developer.
t. a literal hatchling Rust fag (currently learning!), documentation writer and program packager :story:
orowith2os is the troon that kicked off the whole Vaxry debacle. Definitely a proto-cow.
 
It blew me away when I found out that Wayland isn't even a proper display server.
I honestly didn't look incredibly long at the spec and it was a while ago (after I had tried the example wayland window manager actually, name escapes me right now) but that was basically it. The idea behind wayland seems to be to improve on X shortcomings by basically implementing nothing of what X does and declaring "not my problem, you figure it out". Not addressing specific, real world, absolutely common problems is not exactly solving them. From seeing the devs interact with the public I got the feeling they had some serious sunk cost fallacy going on and this was about a decade ago, I can only imagine how they're now but this thread seems to give good examples.

Even though it's not really what wayland does directly, if you wanna assume evil you could also interpret it as another systemd-esque approach of making development of a part of the system so difficult (by having to implement so much yourself) that you once again are beholden to big companies like red hat to deliver a centralized solution for an important part of the system which will end up being so big and complex and interwoven with everything and also fast-moving that it might as well be proprietary. I do admit it is kind of a stretch though. I think it's just poorly thought out. Fact is though that you can write a X window manager on a weekend and that the one I use is so simple that I write patches for it myself. X itself is also often said to be overtly complex and in parts that is true but compared to a lot of current day software, objectively it really isn't. Just has some cruft that would need to be cut away.
 
Last edited:
Re-iterating something I said either in this thread or another. GNOME and Wayland are fucking cancer. And so is everyone who vehemently defends the two when problems arise with applications developed by people that didn't drink the kool-aid. GNOME is a fucking joke. Years pass and you can't fucking drag and drop from an archive manager. Their response? Oh, these archive managers need to get with the program and use GTK4. And you'll see absolutely slime people coping about not having drag and drop. "Oh, a workaround is just to use "Extract files" and select the folder you want the file to go." Or you know I could just install basically any other DE and drag and drop just works. Any self-respecting code maintainers should do themselves a favor and flat out REFUSE to adapt their code for anything GNOME or Wayland. Especially when the people complaining are fucking assholes that can't do anything on their own. That's a YOU problem. Nigger, you are talking to someone that works on a PS2 emulator. They want to work on a PS2 emulator. Not a fucking headache trying to make the program not shit itself on Linux when the end user can just use X.Org when they log in. It's basically troon logic. Respect my pronouns and gender identity! Or else! I identify as a FAGGOT Wayland user and my pronouns are shit/shat. If Wayland was actually you know, DEVELOPED and didn't feel like it was in an eternal, buggy alpha stage, I could understand, but the shit is broken by design. Missing features are hand-waved as good things, or alleging it's for security. I've been using Linux off and on for the past decade, and at no point of that span of time have I ever been convinced the stick with Wayland when it was set as the default on any distro I was trying at the time. There was ALWAYS problems. If I want my video games to work and fullscreen applications to not absolutely shit the bed, I'll use X.org. If I want to set custom resolution timings to fix an issue where my GPU is constantly at 100% memory clock at 144hz because the vblank is too small, I'll use X.org.
 
I've always used XFCE and, despite it being ugly as shit, I have never had the problems other people have had. XFCE truly is the "men really live like this" meme of DEs.
And I don't want it in any way whatsoever. It just works.
An ideal solution in my opinion would be some kind of generic server for all forms of input, audio, and video that can be piped wherever it needs to be, usually into speakers, out of peripherals, and onto screens. That way you could set up whatever architecture you need, at the cost of lots of implementation complexity.
Probably this project might interest you: it looks really promising.
 
I had no idea how lolcowy the Wayland devs are.

They see themselves as pioneers creating a new future standard, but they believe the app&user should not have the final say in where they want windows to be positioned. Why are so many devs commies? And trannies? And gay faggots?

I grew up during the time where mostly tech nerds were on the internet and everything was edgy, how did that create this? Did everyone just call each other a retard so much it manifested into the real world?
 
They see themselves as pioneers creating a new future standard, but they believe the app&user should not have the final say in where they want windows to be positioned.
They believe this should be the task of the window manager, rather than of the protocol. They're not wrong in that, DEs like KDE have features like window zones which need so much more to work right, and tiling WMs don't need any of this, so it does make sense to keep complete implementations of this stuff out of the wayland spec. The issue is just that they take the GNOME stance and rather than make it easy for DEs using wayland to implement new features, they just say "not our problem, fuck off" and then try to plug up whatever hole the DE devs were using to make things work in the first place.
Gnome keeps trying to be Mac and so it inevitably copies the faggotry filled self assured smugness of Mac.
Haha, no. GNOME's problem is devs insist on making their own perfect desktop and then removing any opportunity for plugins to let users customise it to fit their wants instead. Mac is nothing like that. A Mac is a super polished experience out of the box, but Apple put up zero barriers against modifying it with plugins and third party stuff. Yeah macOS is missing some nice-to-haves like window zones, and I actually agree that's pretty bad, but it's an easy matter to install Moom or Rectangle. GNOME on the other hand would look at what Moom or Rectangle was doing behind the scenes, declare that this is "not intended functionality", and block them from working.
 
X itself is also often said to be overtly complex and in parts that is true but compared to a lot of current day software, objectively it really isn't. Just has some cruft that would need to be cut away.
The problem was the client/server architecture which you can't remove without rewriting the whole thing, so they said fuck it let's just make Wayland instead.

It was originally intended for embedded devices where all the excess of X wasn't needed and took way too many resources when you were trying to run it on jalopy hardware. Then it got way out of hand.
 
  • Informative
Reactions: Doctor Neo Cortex
They believe this should be the task of the window manager, rather than of the protocol. They're not wrong in that, DEs like KDE have features like window zones which need so much more to work right, and tiling WMs don't need any of this, so it does make sense to keep complete implementations of this stuff out of the wayland spec. The issue is just that they take the GNOME stance and rather than make it easy for DEs using wayland to implement new features, they just say "not our problem, fuck off" and then try to plug up whatever hole the DE devs were using to make things work in the first place.

Haha, no. GNOME's problem is devs insist on making their own perfect desktop and then removing any opportunity for plugins to let users customise it to fit their wants instead. Mac is nothing like that. A Mac is a super polished experience out of the box, but Apple put up zero barriers against modifying it with plugins and third party stuff. Yeah macOS is missing some nice-to-haves like window zones, and I actually agree that's pretty bad, but it's an easy matter to install Moom or Rectangle. GNOME on the other hand would look at what Moom or Rectangle was doing behind the scenes, declare that this is "not intended functionality", and block them from working.
Seemed to me like Wayland started off with only the bare minimum, which resulted in a lot of shitflinging to get anything standardized, and then GNOME will fuck off to make their own implementation to exacerbate the clusterfuck. Of course, that implies that the pulls get merged instead of being told to fuck off (absolute window positioning).
 
Seemed to me like Wayland started off with only the bare minimum, which resulted in a lot of shitflinging to get anything standardized, and then GNOME will fuck off to make their own implementation to exacerbate the clusterfuck. Of course, that implies that the pulls get merged instead of being told to fuck off (absolute window positioning).
The problem is that Wayland is a deeply flawed technology to be used on a desktop system while it could be useful on a kiosk or mobile platform. One thing that I always asked myself if why couldn't the X11 developers use these 15 years (yes Wayland is that old) to actually improve an existing technology instead of starting from scratch?
I mean... The Wayland creators are the same people that created Xorg as far as I know (correct me if I am wrong)
 
Back