With the transition from X11 to Wayland in full swing, from popular distributions removing X11 sessions altogether and the two major desktop environments planning for the removal of X11 support as well, there’s a ton of questions people are dealing with. Both the KDE and GNOME project published detailed blog posts about the matter.
First, KDE’s Nathan Graham makes it very clear that KDE Plasma’s X11 sessions continues to be maintained. This means KDE Plasma will continue to work on X11, major bugs in the session (e.g. can’t log in) will be fixed, and really bad regressions in the session may eventually be fixed. That being said, minor bugs will probably not be fixed unless someone pays for it, and new features in the X11 session will not happen at all, unless someone pays for it.
KDE currently has no time frame for when X11 support will be dropped from KDE Plasma, and Graham doesn’t expect it to happen within the next two years. The KDE project maintains a list of known significant issues with KDE Plasma on Wayland, and KDE plans on addressing everything on that list before removing X11 support. Graham notes that in the end, dropping X11 support from KDE Plasma is mostly up to distributions, as it wouldn’t make any sense to drop it if distributions aren’t on board. At the moment, about 70-80% of KDE Plasma users are using Wayland, he notes.
On the GNOME side of things, Jordan Petridis also detailed GNOME’s position on Wayland and X11. GNOME will be disabling the X11 session in GNOME 49, with a full removal of the X11 code in GNOME 50. This won’t break any X11 applications (on either GNOME or KDE), since even if they don’t have a Wayland backend, they’ll run just fine using XWayland, which is an X server running on top of Wayland. XWayland isn’t going anywhere any time soon.
According to Petridis, the Wayland session is as functional as the X11 session, and “in plenty of cases a lot more capable and efficient”. He further adds that “there’s some niche workflows that are only possible on X11, but there isn’t any functionality regression”. Basically, if you’re using your spacebar as a heater, you might run into problems.
As for accessibility, Wayland is actually doing pretty great.
There has been a lot of concerned trolling and misinformation specifically around this topic sadly from people that don’t care about it and have been abusing the discourse as a straw man argument. Drowning all the people that rely on it and need to be heard. Thankfully Aaron of fireborn fame wrote recently a blogpost talking about all this in detail and clearing up misconceptions.
↫ Jordan Petridis
Finally, Petridis summarises why the Linux desktop world is moving to Wayland:
No, the Xorg Server is still very much maintained, however its development is halted. It still receives occasional bugfixes and there are timely security releases when needed.
The common sentiment, shared among Xorg, Graphics, Kernel, Platform and Application developers is that any future development is a dead-end and shortcomings can’t be addressed without breaking X11. That’s why the majority of Xorg developers moved on to make a new, separate, thing: Wayland.
↫ Jordan Petridis
This pill is so hard to swallow for some people that they go full bananas and start seeing red hats and Illuminati symbols everywhere, losing their minds and spiraling deep into ludicrous conspiracy theories. The truth of the matter is, however, blatantly banal: the people developing X.org realised long ago that meaningfully improving it would irrevocably break it, and as such they developed something new so they wouldn’t have to break X11. That’s it.
X.org will continue to exist and live on in its maintained state, and desktops relying on it will continue to function. If you want to keep using GNOME and KDE, though, you’ll have to drop X11, because the kinds of features and improvements these desktops want to deliver are not possible without breaking X11. Would you want an X11 that’s broken for everyone, or an X11 that keeps working as-is, while those that want to move on do so somewhere else?
One of these days Thom will post an article that doesn’t have any X11 bashing.
RumblePony,
When it’s retro, but not before. Because Thom has a soft spot for retro 🙂
I have been told many times to have a blade sharp tongue and I certainly enjoy a good bashing and banter. But can’t he get his story consistent at least?! How do those two lines make a narrative:
vs.
We certainly do need different opinions, preferences and educated disputes. But PLEASE lets step it up a bit here.
The Plasma X11 session is going away, there’s just no firm timeline yet, as the article and links clearly state. The article also clearly states and links to the exact few remaining issues KDE is working to fix before deprecating their X11 session. So yes, if you wish continue using KDE over the coming years, you’re going to have to switch to Wayland.
Instead of your usual personal attacks, you could also just work on your reading comprehension.
Guess, what? We are all going to die, there is just no confirmed timeline yet!
I predict that X11 will stay longer than KDE. Mark my words.
Once I was young, a senior gave me a very valuable advise: Communication is the responsibility of the sender.
Jesus! It’s just a piece of software. Get over it, you’ll live longer.
What the hell? Thom, please whatever substance you are enjoying: “small small” and responsible! You get people worried over here.
Again, reading comprehension. Improving X11 beyond the state it’s in now will mean breaking X11. That’s not my opinion; those are the words of the very people developing X.org. All of this is clearly stated in the articles and links.
When I run out of LTS releases, we’ll see whether the following two needs are met or whether I’ll have to do something like stacking rootful XWayland on top of a kiosk-oriented Wayland compositor and going back to something like Openbox for my WM:
1. More than just Qt 6 having upstreamed https://invent.kde.org/plasma/kwin/-/wikis/Restarting for compositor crash recovery. (With compositing turned off, I can run X11 for months without crashing out of my session or things starting to glitch.)
2. https://github.com/xremap/xremap plus KWin Scripts turning out to work as well as I anticipate it will as a replacement for the scripts I write to do things like “If Firefox has focus, block it from seeing Ctrl+Q keypresses” (thankfully, no longer necessary) or “If Thunderbird’s WM_WINDOW_ROLE=’3pane’ window has focus, block it from seeing Esc keypresses”. (XRemap DOES say it works with KWin Wayland and, under X11, I like how well it works for what I’m already using it for.)
…so it’s really mainly a question of “will someone come up with an alternative Flatpak GNOME Runtime that incorporates GTK patches so compositor crashes don’t kill my Inkscape despite GNOME ego declaring compositor crash recovery unnecessary and a WONTFIX?”)
This may not be what you are looking for but, if you like Openbox, check-out Labwc on Wayland.
Once Wayland issues with font rendering and with remote desktops are ironed out, I’ll be more than happy to use it on Linux. On OpenBSD I’ll just as happily continue using Xenocara (their fork of X11) as long as it exists, and there’s also Arcan to keep an eye on though its development seems to be glacial.
On Corporate desktops there are still millions of Java Swing application, depending on X11. They do not work flawless with XWayland and Swing/Wayland support is just a POC so far. I don’t see X11 dying out soon. (Remember, Fortran and Cobol still exist for a reason, legends never die!)
I’m totally fine with wayland. I’ve used it for a month solid last year and it rendered the screen the same as Xorg.
But there’s no way I’ll use Gnome or KDE. I use a minimal tiling window manager called DWM where everything is just a key press away. Gnome and KDE are kind of the opposite of minimal and tiling and keyboard driven.
When DWM switches to wayland, I’ll be more than happy to make the switch. But until then, Xorg isn’t going anywhere.
Here’s Wayland’s dwm: https://codeberg.org/dwl/dwl
I, for one, have settled on Sway years ago and everything work perfectly fine for me. I’m still a bit sad to have lost AwesomeWM, but such is life. (I switched because a game was crashing on AwesomeWM/X11, but was stable on Sway/Wayland.)
Yes, I should give DWL a try. Probably this summer sometime.
Somebody else recommended DWL as a DWM replacement. That makes sense because it is supposed to explicitly be that.
However, if you like tiling window managers, check out Niri on Wayland:
https://github.com/YaLTeR/niri
It is quite easy to call others with pejorative adjectives (e.g. conspiracy inventors). But there are reasons for some people not to be happy with the way this transition to Wayland is going.
bemcl,
This is my take too. Don’t blame users, users aren’t the problem! Xwayland dropped some features that many of us needed and it’s responsible for much of the friction. Wayland needs to fill in the missing gaps, if a feature only works with certain hardware and/or certain desktops, then wayland remains incomplete. At least support seems to be improving. Too often though developers won’t do enough to work with users at the begging when it’s most important. This sets the whole project back by years to find ways to deliver on unmet user requirements afterwards.
So far, nobody could answer the most simple question: IF I installed Wayland today, what would I benefit from right now (on user level)? What positive difference would I notice? Just for a normal user, nothing fancy like X-forwarding or stuff or Java Swing UIs.
– Different refresh rates when using multiple monitors (e.g. because your laptop has a 60 Hz display, you are no longer limited to 60 Hz on your external display)
– Different scaling for different monitors (e.g. just because your laptop screen needs 2x, you don’t need to also use 2x on your external display)
– Proper HiDPI support
– Color management, e.g. for DCI-P3 which almost all modern displays have
– HDR
– Application sandboxing finally works (in X11, every app can control every other app, breaking any sandbox)
To name just a few… But I’m sure someone will shout that they need none of those and hence Xorg is better and everybody who needs any of those is an idiot and yada yada and more copium.
Thank you for this list, it is indeed a solid basis for a discussion and I fully understand that anyone in need for those features will be very interested in Wayland.
Although I do not understand where the “idiot” comes from and why it was needed here.
Let me give you a simple analogy: In German I am actually very interested in Ice scrappers and bought at least 2 for each car. In Thailand, not so much and your business of selling those tools would met with less interest. Nothing wrong about it, both points of views are perfectly valid to me. No “idiot” needed, was it?
Seems I cannot reply due to too much nesting, so I’ll respond to the parent instead.
The idiot wasn’t me calling anyone an idiot, this was me saying how these discussions usually go: “Wayland can do nothing X11 can” “It can do all these things X11 can’t do” “Nobody needs that” “I do” “Then you’re an idiot”.
All of those are great reasons to switch to Wayland and are welcome improvements over Xorg. Unfortunately I rely on remote desktop which is still not 100% working (though Krdc at least is working for the most part and is usable in Wayland now). Also Plasma and Wayland causes font rendering issues for me, so I have no choice but to fall back to Xorg to save from getting a headache. Once those issues are ironed out I’m ready to move to Plasma on Wayland full time.
With that said, there are those folks who have other reasons for wanting to stay with X11/Xorg, and I feel like their complaints are constantly ignored by the Wayland evangelists who insist that “just try another compositor bro!” is a valid response to any issue with Wayland in general.
Wayland is past the “half baked” point in its development cycle, but it’s not 100% ready for everyone.
@Morgan
Do you mind sharing what distro and KDE version you are using? I am quite curious who is having issues and who is not.
Some of the issues described are so surprising to me. I had even started assuming that all these bad experiences must be on Debian. Espaecially after a recent post attracted a comment that Wayland did not work because of a bug in the pre-alpha release of Q4OS (a bug already fixed in that alpha before the post).
But then somebody claimed a bad experience in Void which should be pretty up-to-date. So, now I am super curious.
Plasma is second-classing X11 because Plasma 6 on Wayland works better than KDE 5 on X11. At least, that is what most people are experiencing. So, I am really curious about the environments where this is apparently not the case.
@LeFantome:
It’s Plasma 6.3.3 (latest available in the repo) on Void Linux. As I said, Krdc is actually working fine now, I can use it under Wayland with no glitches or issues; this wasn’t the case a few months ago. Remmina is still glitchy under Wayland, giving me either a black screen or a garbled screen for the remote device. NoMachine (which I use to connect to work computers from home) works fine under Wayland.
The font rendering issue still exists, and it’s pretty nefarious. Any time I scroll text in Kate or LibreWolf or Konsole or really any scrollable text, it will “split” random lines of text. It’s as if someone took a laser and dragged it across the middle of the line of text, then shifted the top half to the right by a pixel or two. Scrolling again might make the splits disappear or it might make them move to another random set of lines. This doesn’t happen in X11, so it’s definitely a Wayland issue. It was happening as recently as a few weeks ago.
To put it simply… X works, Wayland does not. Likely will never be on par with X. Just like what happened with the massive downgrade from KDE3.x to Kde4+ (or from Gnome 2.x to gnome 3+)
I would add Wayland is sensible faster and, more important to me, its responsiveness is kilometers away from X11. For a desktop* user, those are very important.
*I mean, for browsing, playing videos, music, coding/editing text, etc. I do understand a desktop can be used for stuff that is more focused into a terminal or single app too, so I am talking desktop a more generalistic and casual user 😉
protomank,
I tested this myself around two years ago when someone suggested wayland was faster. I measured the frame rates I could push under wayland and X11, and for me X11 was actually marginally faster. I’m not saying this is universal across hardware or that it’s a static situation where wayland can’t improve. However you can’t really rely on the echo chamber to be true, you need to test your hardware. Many point to the X11 socket overhead but don’t realize that most accelerated applications bypass that.
At least for me the difference was small enough that it wasn’t really a big factor either way.
To me one absolute showstopper with Plasma Wayland,
No usable session restore.
With 10 desktops (or activities) and 20+ windows (many with multiple tabs) in a session having to manually restore (or even rearrange) them is not an option.
Showstopper also for me, like remote desktop.
It is a bad modern habit: you do a new thing that misses important features of old thing and then wonder why people do not run to use the newer one.
@James-T
Wayland has the xdg-session-v1 protocol. It is implemented in Plasma 6.4.
Yeah, I was going to say the same. I’m using KDE 6.4 and session restore is just fine.
What if wayland simply worked and supported stuff that already works on X11? People would switch overnight and forgot X11 even existed. Maybe hardware acceleration hardly works on wayland and screen glitches when switching fullscreen modes, but in exchange remote access and accessibility doesn’t work.
It’s as if the compositor, the windowing system, and the desktop environment are so tightly coupled to each other that separating them doesn’t make sense.
Sure, you can separate them, but then you have the desktop environment being restricted by what is offered by the compositor and windowing system underneath (that the desktop environment can’t change because they are “frozen” and other desktop environments rely on them too).
Even the success that X11 had in Unix workstations, it was because it was defacto coupled to Motiff and CDE (or modified versions of CDE like IRIX Magic Desktop) which were designed with the limitations of X11 in mind.
Apple discovered this back when they were designing what would eventually become MacOS X, so they wisely decided to keep the whole stack integrated, and it has served them well because they have been effortlessly changing the MacOS X desktop environment to their will without being restricted by the layers underneath.
That said, that will never happen in Desktop Linux land, so just give me a minimal mostly-modern compositor like Wayland without a windowing system attached, so at least the desktop environments can bring their own windowing system and are restricted only by the mostly-modern compositor (Wayland).
Well, isn’t that the whole point of Linux?
Early 2000s, late G4 era, I migrated from Windows to the Mac, and it was awesome. UNIX power with apps like Photoshop. After Mojave, I became very disappointed with the direction Apple was taking the Mac, so I migrated.
I’ve always lived a parallel life with Linux. After I became disappointed with the way Linux was going, I seamlessly switched to FreeBSD. I also went from GNOME to KDE and, after that also became too eye-candish to my taste, I switched to openbox, with all the applications I use perfectly surviving changes and remaining perfectly useful.
I own and regularly use an Octane with IRIX 6.5.30 and I am not sure your comment about IRIX Magic Desktop being built around the limitations of X11. SGI had its own Xsgi library that they extensively used and you could also use when writing applications to IRIX. As far as I know, the only reason this library worked so well was due to X11’s modularity.
The main risk (again, in my opinion) of everyone going all-in on Wayland is ending up where we are in 20 years. If DEs and application developers keep the discipline of targetting multiple platforms instead of settling for Wayland-systemd, we can avoid having the whole world stuck once again. Highly portable applications are very good for the environment and also to allow us to use computers for longer. It also keeps developers of main infrastructure components (Wayland-systemd) honest, as there would be always a competing environment to benchmark your own decisions regarding features, performance, functionality and security.
This kind of change, for me, is very traumatic. It took me 10 years to migrate away from Aperture. I take photos semi-professionally and the computer has always been a tool for the craft. Whenever someone reinvents the wheel, unless the benefits are crystal-clear and the migration path butter-smooth, migrating is very painful. Scripts, backup schemas, configuration, workflow…
So yea, I once took the weekend to give Wayland a shot, had issues with my color calibration workflow, and that was it. I am not sure when I will have time to try again. I don’t turn the computer on to use Linux, Mac, or Linux. I turn it on to load applications and get work done. If computers were screwdrivers or a drill, we would find the handle changing position every weekend, the buttons getting new labels or suddenly you would find out that what had to be done clockwise now is done counterclockwise.
I thought I could escape this sort of end user bullying by moving to Linux from MS or Mac, but it seems not!
It’s a bit like Devs always pushing or enforcing the latest and greatest GUI updates on me, while threatening a boycott and complete destruction if anybody dares to even alter the serif on the font of their favourite terminal or command line utility! They seem oblivious to the hypocrisy.
I devoutly hope that I never have to use GNOME or KDE Plasma for my daily driver computer.
I write about them both, a few times a year, and every time it reminds me why I do not use them and why I do not want to.
KDE is a baroque mess with so much duplication and redundancy that it’s tragi-comic.
GNOME is an attempt to bring the iPhone UI to the desktop without understand why the iPhone team made the decisions they did.
I hope that the moves to tie GNOME to systemd mean that the BSDs and the systemd-free Linux distros will drop GNOME, because that will help the many other desktops out there.
Probably over 50% of new Linux desktop installs are GNOME. My preferr.ed distro considers GNOME the default.
I do not use GNOME myself.
Most of those installs are coming on distros that use systemd though.
You raise an interesting point. Many of the systemd free distros may be able to get away without shipping GNOME. That would give other desktops more of a chance.
A colleague recently updated his laptop and got Wayland underneath KDE and wasn’t able to log in anymore. The only fix we were able to find was to manually switch it back to Xorg. So KDE may want to first fix Wayland before switching to it as the default.
Wayland in Plasma (not plasma 6.4) has been working flawlessly for me as a daily driver.
Specially the different per screen scaling.
And then how plasma workflow just works and is stable.
I don’t want ever to go back to X11.