Wayback, the recently announced tool that will allow you to run a legacy X11 desktop environment on top of Wayland, has just announced its first release, version 0.1. As the version number implies, there be dragons here, but the developers state some of them already use Wayback on a day-to-day basis. Still, there’s no multi-monitor support yet, quite a few X.org options are just stubs for now, there’s no mouse-locking, and so on.
Since the initial announcement and the first progress report a few weeks ago, Wayback has become an official part of FreeDesktop.org, which indicates the wider desktop Linux community is definitely interested in what Wayback has to offer. It’s also been split into several different parts to mimic X.org’s structure, several distributions have picked it up and packaged it already, and ton more changes have been made.
It definitely seems like Wayback has a good chance of becoming a simpler, more straightforward replacement for X.org, greatly reducing the maintenance burden of Linux distributions. Not having to keep the full legacy X.org stack around alongside Wayland is going to save a lot of people a lot of time.
This is obviously not going to be in a state to satisfy the “Wayland breaks everything” hard-cores, but I just tried it and it works pretty well already. All I did was create a simple .xinitrc to launch fvwm3 but it popped up right away. A few quick tests showed that it reported as X11 and I had no Vulkan support. But glxgears worked as did a few other test programs. If I just sat down at it, I am not sure I would have suspected I was using anything other than fvwm3 on Xorg (although inxi did report Xwayland). This was just on a laptop so only one screen.
I built from source and it build like lightning on a low-end laptop. And the binaries are tiny. Xwayback is 28k and wayback-session is 15k. I mean, it makes sense since it is just a super stripped down Wayland compositor. Xwayland is 2.2 Mb. But I had that already.
wayback-session already runs in Wayland if you just want to run it in a window.
This is what I’m looking forward to the most. One of the great things about Wayland is that it’s tiny and tidy, it’s a perfect fit for constrained devices like Raspberry Pi boards. Using Wayback instead of having to pull in all those Xorg dependencies means things can stay neat and tidy.
This is the kind of thing I always wanted out of Wayland: make it just a piece of plumbing that makes managing hardware/drivers less crufty and leave all of the “good enough” X11 stuff on top. There shouldn’t *be* a lot of “native Wayland” programs–there’s just no need for most of them. A few, sure, where it makes sense and there’s a benefit. Everything else? Leave it as it is. When there’s a compelling Wayland-native alternative that covers all the use cases it can then be trivially launched alongside everything else.
Doing so you only gain compatibility with new drivers but loose all of the other benefits of Wayland, especially security. By design every X client can listen on every other client, by design, on X, it is impossible to prevent keyloggers.
There is good reason to push with the Wayland transition. There will always be use cases that are uncovered and even unknown to the developers. Pushing the transition makes the framework and DE developers invest time they otherwise would not, and makes users of niche use cases complain, therefore bringing said niche use cases to the attention of the developers.
alexvoda,
“Impossible” is too strong a word here. I don’t contest that this is one of the features that is considered insecure, however as a developer it would not be so difficult to add a permission check to make it secure. However for the devs in charge of Xorg, wayland is the answer rather than fixing X11. This has happened with other missing X11 features as well. For example HDR isn’t possible on X11, which is true. But it’s true because Xorg neither accepted nvidia’s HDR patches nor did they propose their own HDR implementation.
https://www.phoronix.com/news/X11-DeepColor-Visual-RFC
The closer one looks, the more instances of this we see with Xorg. I don’t want to be unfair to them, it’s their prerogative to accept/reject features as they see fit. But at the same time it’s difficult not to cite it as a real conflict of interest, especially to X11 users. The organization responsible for X11’s code base only wants bug fixes and doesn’t want X11 to get new feature improvements in order to promote the wayland agenda instead.
I feel X11 is bloated and there’s merit in replacing it. I’m all for progress and I don’t intend to keep running X11 once wayland is working for me. However this type of argument, where we point out that X11 features are missing/broken, comes off as somewhat disingenuous when Xorg have been impeding development of X11 features to promote wayland.
It’s not all that different from microsoft trying to promote metro instead of the classic desktop by actively kneecapping the classic desktop to make the experience worse, which is what MS did with windows 8. Whatever the true merits of metro and the classic desktop were, it would be disingenuous to promote metro to address classic desktop deficiencies that they themselves were causing. I get the same vibe with some of the arguments for wayland over X11.
@sorpigal
> This is the kind of thing I always wanted out of Wayland
What confuses some of us is that you have been able to do this for quite a long time now. Xwayland was introduced in 2014 and always had rootful mode. Cage first appeared in 2019. With just that combo, you could run Cage (a kiosk Wayland compositor) with Xwayland on top.. Tada, a full X desktop on Wayland, featuring whatever X window manager you want. The marketing is better for Wayback, and they are doing more to make it a “drop-in”, replacement (like running the stuff you have in .xinitrc) but Wayback is not new tech. There is a reason it only took them two weeks to go from announcement to first working version.
> There shouldn’t *be* a lot of “native Wayland” programs
Here is what the stack looks like with Wayback:
Hardware -> Wayland compositor (Wayback) -> Xwayland (X server) -> X11 applications
Here is what normal Wayland looks like:
Hardware -> Wayland compositor -> Wayland apps
When you look at it that way, there shouldn’t *be* a lot of “non-native Wayland” programs. Xwayland itself is a native Wayland application. What do the extra layers add? Many regular users will not even use Xwayland two years from now.
That is why many applications are gong Wayland only. By complete accident, the first application I tried to run on Wayback did not work because it is Wayland only (the foot terminal emulator). That is why GUI toolkits like GTK are going Wayland only. That is why Desktop Environments like COSMIC, GNOME, and KDE are going Wayland only. That is why projects like Hyperland, Niri, and Cosmoe are Wayland only. And remember, even the Qt FLTK, GTK3/4, Electron, and WINE apps will be Wayland apps when run on a Wayland compositor, even if they still support X11 when run on an X server. Apps still requiring Xwayland will be Athena toolkit, pre-GTK3, Pre-Qt5, and Motif. How much of that do you run now?
> where it makes sense and there’s a benefit
The transition is in its infancy and Wayland applications already have many benefits. HDR. variable refresh. fractional scaling. sand-boxing, security. no tearing, compositor bypass, simpler design, lower lag, and probably many more.
The list of advantages of Xorg over Wayland shrinks every day. The list of benefits to Wayland grows. And I estimate that over 50% of Linux desktops are Wayland already. Debian Trixie is about to add a few more.
>Wayback has become an official part of FreeDesktop.org, which indicates the wider desktop Linux community is definitely interested in what Wayback has to offer.
Not saying that Wayback isn’t interesting, but that sentence is a non-sequitur as FreeDesktop.org is not a democracy and doesn’t care about what’s popular with users.
Wayback sounds like the sort of thing that Wayland should have had _right from the _start.
ie. make it easy for people to slide from X11 into Wayland without having to completely blow up their existing desktop habits.
You are not alone with that sentiment.