It won and is good.

PiP is still broken and by the way they argued about the implementation it could be the next file picker meme.

I randomly tried GNOME and noticed some of my older games had much smoother animations. now I use dash to panel and abandoned Cinnamon because the performance is so much nicer. I was naive, but now I've seen the light. the principle of having a modern compositor makes a lot more sense than hobbling along nearly 40-year old tech and adapting it to modern needs.

>>108813927
dogshit protocol with okay implementations, as opposed to x11 being an okay protocol with a dogshit implementation (xorg)

It just works for me

>>108813927
>impossible to have autohotkey-esque functionality
i guess there was no use case for automating things on linux

>>108813927
Is wayland just a protocol like tcp/IP and it's upto programs to implement it how they want?

Why did they go with this approach? Doesn't this mean every DE has to implement the exact same shit which would cause unnecessary labor instead of a unified server which all can use?

>>108813927
It's run by idiots who looked at X11 and didn't ask "how can we do things better" because they were fixated on "how can we do things different". Totally failed to learn the lessons.

>>108816400
on all wayland compositors that i know of, you can configure a keybind to run an arbitrary command, so you could just bind to a shell script

>>108813944
fippybippy

>>108816931
in some scenarios yeah. in every compositor i know of THO, an open focused window eats the inputs instead of passing them along to the compositor.

>>108813927
finally started working on my machine when i got rid of my jeetvidia gpu

>>108813927
it's ok but it's got a problems that will never be fixed because waypists at this point have given up on actually trying to make wayland better and instead have taken to sabotaging xorg and xlibre instead kek

>>108817110
what

>>108818922
sounds like simlish

>>108813927
It's one step closer to making desktop Linux secure. Before Wayland there was literally no separation of applications and any of them could just see what the others were doing.

>>108819282
>nooooo I have to run malware locally, I just have to okay?!?

Since pure Wayland doesn't do much at all a giant protocol soup emerged that tries to cover all the missing pieces.
When writing native clients you have to fight your way through a vast labyrinth of boiler plate callback registration and capability check spaghetti before you even can paint a single pixel on the screen.

Wayland is so bad it wouldn't be far fetched to assume it is intentional sabotage.

>>108819772
>Wayland is so bad it wouldn't be far fetched to assume it is intentional sabotage.
What's so bad about it? It just works

>>108819757
Do you understand what a vulnerability is? You don't have to run a malicious executable to get pwned. You can get drive by pwned from viewing and infected image in your image viewer or opening an infected FLAC in your music player.

>>108819813
> What's so bad about it?
I specifically pointed out the reasons. Learn to read.

> It just works
It works perfectly as sabotage.
There are plenty of things that don't work. But even if they would Wayland would still be a dogshit protocol.

>>108819772
>Wayland is so bad it wouldn't be far fetched to assume it is intentional sabotage.
why do you think XLibre got so much hate? (((red hat))) tried to permanently kill Xorg by refusing maintainers then claiming no one wanted to do it.

>>108813927
I like it's push towards combining the compositor and window manager (in the tiling window manager world) as it allows stuff like Hyprland/MangoWC/Niri to exist, but it also killed the "make a few something in thousand lines of code" era for TWMs. River is trying to fill that gap, but I haven't tried it much (and when I did, I had issues). Tbh I'm looking forward to Phoenix if it ever becomes something since it keeps the X11 protocol while working more like Wayland in practice.

File: 1754879135093409.png (136 KB, 499x499)

136 KB PNG

>>108813927
It's fine for trivial applications and tranny screen autism, but it will never be the future of Gaming (the only thing that actually matters with computers) because it refuses to acknowledge that programs might want to interact with eachother, so overlays etc are only possible with retarded DE-specific workarounds. Mark my words, in 10 years some arch-based xlibre distro will be the king of gaming.

>>108817113
You need a system that hooks directly into the input source (i.e evdev).

Ydotool and Hawck work even if you were set at a TTY with no desktop session running.

https://github.com/ReimuNotMoe/ydotool
https://github.com/snyball/hawck

>>108820399
>because it refuses to acknowledge that programs might want to interact with eachother
This is a great feature DOE

>>108820399
>Chud wants his desktop filled with malware and adware popups / overlays
Whyyy?

>>108818080
This is the only reason I dont recommend Linux to my friends. Almost all of them got scammed by njudea, and recommending an OS without wayland to someone with decent computing power coming from windows is just malicious.

File: 1764255202546396.png (233 KB, 563x542)

233 KB PNG

>>108820438
>>108820454
Userland is still far from secure. Fucking the user in the ass doesn't help much. Use containers if you are actually concerned instead of controlling the user's behaviour. But I'm not here to tell you what to do. We will see which server Gamers are using in 2050.

>>108813927
Not much. Did some things right and others wrong. The goal was to make it a lot more modular and simple and they succeeded on that, but at the same time created a lot of fragmentation, which is being solved with specs, but creates some temporary chaos.

What it does better than X11:
- More modular allowing different implementations or enable/disable features
- Better performance by directly accessing the kernel
- More abstract allowing for devices other than screens like VR
- Better security by applications containing their own state, for example, copy paste

What did poorly:
- Missing features took a lot of time to materialize
- Created fragmentation specially with experimental features
- Harder to create new WMs
- Nvidia didn't want to play along until the 5000 series

>>108820457
You can get away with the 5000 series but below that it's a mess

>>108820515
A container doesn't help you with graphical software. If the display server allows it to make popups / overlays and it has a connection to the display server than it can do whatever the fuck it wants as long as it has access to the socket.

Only solution is to run a nested display server but that comes with more overhead. Better to just forbid this crap in the first place. Nobody needs your shitty overlays.

>>108820515
Other than what Anon said, you can just force x11 to any overlay app if you want to be a tranny about it.

I don't know what it does
seems better than X11 so

>>108813944
/thread

>>108820565
It's OK, just keep telling users what they are allowed to do. Let's see what happens.

>>108820630
Nothing will happen because nobody wants these shitty overlays. They look shit, act shit and are shit.

People actually prefer tighter integration with their desktop.

>>108820644
tighter integration? what does that mean?

>>108820673
It means people to go raw without a condom. I'm not the person you asked but that's what it means.

>>108820673
>so overlays etc are only possible with retarded DE-specific workarounds
People prefer those "DE-specific workarounds" because it actually integrates with the desktop properly that way instead of being a pile of shit that looks like a pile of shit and acts like a pile of shit.

>>108820687
Let's be honest, the large majority of users only use flatpak because it has a store and allows to install the latest software in traditional distros. There's very few people that actually like the container aspect of it.

I'm agreeing with you, just adding more context. In arch, cachyos, nixos, guix, and all the other rolling release friends, the flatpak usage is a lot lower.

>>108820707
>flatpak
NTA but what flatpak has to do anything with this?

>>108820362
>I like it's push towards combining the compositor and window manager
I sure love making the linux desktop both harder to implement and more fragmented!

>>108820749
It has to do with the original opinion that containers are better for security than whatever Wayland is doing with portals and overlays.

>>108820431
based, usability by nuking tranny wayland security theater

>>108820548
>- More abstract
opposite is true
>- Better security by applications containing their own state, for example, copy paste
this is cancer

I can only opine regarding implementations as I don't understand the protocol innards (probably because I'm not writing my own implementation)
kwin is good, cosmic seems fine too, I don't miss anything from x11

i will not use compositors

>>108813927
vrr and hdr just work.

let x11 die already, it served its purpose but its old and clunky now.

remember when there were daily wayland hate threads here?
got real quiet

>>108821364
Xorg chads don't need to post Wayland hate threads because their system isn't constantly breaking.
Most hate threads come from people using Wayland breaking constantly.

>>108820918
You don't know what more abstract means? Window positioning in X11 assumes you're using a desktop, with 0+0 coordinates, in Wayland there's no such assumption, it's all relative. Wayland is more abstract in that sense.

So you agree with Microsoft that storing passwords in plain text during the entire session of Edge is good and that it's fine if other local users can read memory and see your passwords? Wayland protects against that automatically.

>>108813961
Works on my machine

>>108821545
wayland has hyper specific protocols that only allow for specific usecases. x11 has generic building blocks that the implementation can do whatever with, so yes wayland is less abstract.

usecase destroying security theater is cancer, yes

>>108814792
>had much smoother animations
>performance is so much nicer
>I was naive
happens to everyone, switched over this year myself

>>108821573
Give me an example

>>108821573
I think you're confusing vague with abstract. Abstract doesn't mean unspecified, it means you have a specification and you can have multiple implementations that match it, hiding the complexity. The crucial different in technology is you must have a well defined specification, otherwise it's just a wishlist that is incompatible between systems. Or a poor abstraction if you will.

Having specific protocols doesn't mean less abstract in programming.

>>108820548
You also can't run it without compositor. It's deal breaker for me.

>>108822233
You're a retard that doesn't understand what a compositor is or does. When you hear compositor you think of what the compositor is under X11 that adds blur and transparency and introduces input latency and lag due to how shit the Xorg Server is.

That's not what a Wayland compositor is. The "compositor" just means it does the same role as what the Xorg Server does. You don't necessarily need all of the fancy effects and vsync, etc. In fact the compositor practically switches off entirely when full screen applications are running.

>>108822504
What OP meant is that on Wayland you have two mandatory buffer swaps and therefore clients can't directly render to the front buffer.
On uncomposited X11 this is possible.
Also Wayland is a very opaque protocol. When you submit your surface you have no idea how, when and where it is actually rendered.

>>108822538
>Also Wayland is a very opaque protocol. When you submit your surface you have no idea how, when and where it is actually rendered.
Doesn't Presentation Time Feedback solve that?
https://wayland.app/protocols/presentation-time

I'm not sure what Xorg does but afaik presentation time gives clients accurate timing information about when the last commit happened and when to submit the next frame, etc so they can ensure perfect synchronisation of frames and that nothing gets dropped, etc.

>>108822564
> presentation-time
Like so many things in Wayland, this was introduced as an afterthought and is now part of the protocol soup.
Note that even though this protocol is allegedly "stable" compositors still can choose not to support it. This means if you want to use it in a client it also becomes part of the capability checking spaghetti.
Wayland is such an non-ergonomic piece of crap. It belongs into the trash.

>>108822611
Or you can just assume that it works on the two major desktop environments and if someone runs your app on an experimental compositor like Muffin (Cinnamon's Wayland compositor which supports fuck all because it's based on an older version of Mutter) then they can go fuck themselves.

>>108822632
No you can't simply assume. Your program will segfault while registering the callback if you don't query the registry first. Good luck dealing with obscure bug reports.
The fact that something as essential as presentation-time is not baked in and you have to do a weird complicated protocol dance for such basic and essential things shows how far removed the protocol designers were from reality.
Wayland is absolute utter trash. And inventing 50 more protocols on top of it won't change that. It's rotten to core.

>>108822788
Literally nobody cares if you segfault on an obscure experimental compositor. You can just tell them to use a real desktop and go away and stop bothering you.

>>108822811
Even if you do did add a workaround it'd probably run like crap without accurate feedback to drive your rendering loop anyway.
>Don't use experimental compositors
>Use a real desktop

>>108813927
it's ok, it works fine
not like you have a choice in the matter

>>108822811
> You can just tell them to use a real desktop
I already told them to use X11.

>>108822868
Good. Better than an experimental compositor. Wayland is actively harmful when the compositor developers half-ass the implementation.

>>108822868
X11 is a protocol, not a desktop

>>108822538
Man you anons find the dumbest excuses all the time. Double buffering has been a standard in graphics since forever. There's no graphics without double or even triple buffering today.

>>108822233
A compositor is just a layer in rendering to show graphics in the correct order. If you didn't have compositing windows would render in the wrong order glitching all over the place. Who decides which window to render first? The compositor. X11 also does compositing, whether you call it that or not.

>>108823282
Xterm does not use double buffering. As a result it still the best latency values of all terminals to this day.

>>108823304
> X11 also does compositing, whether you call it that or not.
No it does not.
It draws directly to the front buffer using hardware accelerated mask and blit operations.
That's why uncomposited X11 can't do transparency.
You cal look that up in the ddx drivers yourself.

>>108822811
> Literally nobody cares if you segfault
t. the quintessential Waylander

>>108823399
The original X11 used stacking but any modern version of Xorg uses compositing. Stacking is not a great idea because if you mess up the order of window drawing you get flickering, because Windows were rendered in the wrong order. I don't understand why you want this. Pick any game, it doesn't matter, all of them have double buffering.

One thing Wayland does differently however is that each application has a double buffer, not just the screen. That is a bit more expensive but guaranteed no flickering.

Can't do HDR without it
Having problems with the steam controller with it

>>108824406
Did you miss the part where every single compositor works and does not segfault except your out-of-date experimental compositor by a two developers that don't even know what they're doing and would rather work on a screensaver application than actually fix their compositor?

>>108824541
Sometimes Wayland even does triple buffering for slower GPUs:
https://www.phoronix.com/news/GNOME-48-Triple-Buffering
https://www.phoronix.com/news/KDE-KWin-Triple-Buffering-MR

>>108813927
freedesktop is on a mission to destroy linux
nothing has done more damage than gnome

>>108824800
Yes. You can disable it though in some compositors and in gaming. I don't see why you would want to do that in desktop but ok. Any excuse to trash Wayland I suppose.

>>108824921
Read the threads for the WHY they have that. On some shit potato systems they can't handle rendering in real time so they maintain a dynamic triple buffer. This actually results in better performance, you wouldn't want to disable this and get worse performance.

I still cant have an always on top always toggle-able drop down terminal.

>>108824952
Yakuake works fine on KDE Plasma

>>108824972
Also, Yakuake is working on LayerShell support too which might interest you:
https://invent.kde.org/utilities/yakuake/-/merge_requests/99

This will mean it won't rely on private Plasma Wayland methods anymore and will work on Wlroots compositors and other compositors that implement layer shell.