After two years of intense development, the third major Linux desktop environment has released a new version: Xfce 4.20 is here. The major focus of this release cycle was getting Xfce ready for Wayland, and they’ve achieved quite a bit of that goal, but support for it is still experimental.
Thanks to Brian and Gaël almost all Xfce components are able to run on Wayland windowing, while still keeping support for X11 windowing.
This major effort was achieved by abstracting away any X11/Wayland windowing specific calls and making use of Wayland/Wlroots protocols. A whole new Xfce library, “libxfce4windowing” was introduced during that process. XWayland will not be required to run any of the ported Xfce components.
↫ Xfce development team
A major gap in Xfce’s Wayland support is the fact that Xfwm4 has not been ported to Wayland yet, so the team suggests using Labwc or Wayfire instead if you want to dive into using Xfce on Wayland. While there are plans to port Xfwm4 over to Wayland, this requires a major restructuring and they’re not going to set any timelines or expectations for when this will be completed. Regardless, this is an excellent achievement and solid progress for Xfce on Wayland, which is pretty much a requirement for Xfce (and other desktop environments) te remain relevant going forward.
Of course, while Wayland is a major focus this release, there’s a lot more here, too – and that’s not doing the Xfce developers justice. Xfce 4.20 comes packed with so many new features, enchancements, and bug fixes across the board that I have no idea where to start. I like the large number of changes to Thunar, like the ability to use symoblic icons in the sidebar, optimising it for small window sizes, automatically opening folders when dragging and dropping, and so much more. They’ve also done another pass to update any remaining icons not working well on HiDPI displays, removing any instances where you’d encounter fuzzy icons.
I can’t wait to give Xfce 4.20 a go once it lands in Fedora Xfce.
First congratulations on this achievement, but then again another FOSS project with sparse recurses basically forced to spend years on adding Wayland support, result still considered to be experimental at best. Sorry but this ended up being a turd, the whole Wayland thing. We need to start criticizing it instead of perceiving this as some sort of success, for parties involved to get their act together instead of believing they are doing a good job. At minimum somebody needs to take responsibility again, to offer a common Wayland compositor and for that library to be used in most GNU/Linux distributions and in Android. Without achieving that forget about Wayland ever really being a thing. We can waste another 20 years on Wayland and then after turning 40 years old, we can call it obsolete. And invent something new. Modern and picture perfect, give me a break.
Wayland may truly well be The Way Forward, but if this is the case, it’s been a long and slow road for many with many miles yet to go. Will check back on this when given a reason. As for creating a proper toolkit for WM’s and DE’s to build from? I refer everyone to xkcd 927: Now there are n+1 competing standards ; – )
Currently i don’t see much oppressing problems in other areas beyond maybe two. That is common display server/compositor and common packaging format, fragmentation in this two areas is causing too much wasted effort. So far Wayland/Snap/Flatpak/AppImage are failing to resolve it.
I kind of agree. If the same collective effort (and money) was spent on tidying up X11, we’d have the same robust feature set with a more maintable code base. Instead we’ve moved over for limited end user gain and substantial effort across the ecosystem.
Geck,
Wayland’s progress was rocky, and I wouldn’t deny there were self inflicted problems, but thankfully most of that’s behind us. And X11 isn’t getting any less bloated or antiquated, a replacement was justified even if project wayland got off on the wrong foot. I dislike certain aspects of wayland, such as desktop compositors having to reimpliment the same features across desktop environments, the barriers to using wayland have been decreasing or are even gone for most users, meanwhile it will become increasingly evident that X11 has reached EOL as applications get ported to wayland and stop supporting X11.
Personally I am waiting for the newest fixes to make their way to Debian stable, but once they address my reasons for holding out, I’m going to switch too because my reasons for sticking with X11 have been purely pragmatic and not ideological. Regular users don’t have an ideological connection to X11 either, they just want something that works,
When saying most of that is behind us. Is it? We now have a viable solution for GNU/Linux desktop stack and on mobile for Android, utilizing Wayland? Similar to what PipeWire does for audio? Or is it all still more of a pipe dream?
Geck,
I know you want wayland to fail for your own ideological reasons. But normal users don’t really care about that, they just want a desktop that works. I certainly did criticize wayland for dragging their feet on some important use cases, but personally I will get over it as they fix things. It’s not a pipe dream at all. While I readily accept that you may have edge cases that don’t work for you (like I did), you should be honest and admit that many users find it already works for them today and they really don’t have a reason to boycott wayland.
Why would i want for Wayland to fail due to the ideological reasons? What reasons would that be? Users? Xfce users? Lets instead talk about developers first? Developers caught in this uphill battle and only after lets discuss users that do or don’t care.
Geck,
I’m pretty sure you’d rather have wayland fail because the alternative means X11 will continue erode ground to wayland. But if I am wrong and you don’t mind X11 giving ground to wayland, then let me know.
You haven’t been clear about what developers your talking about, but most application developers use higher level toolkits that abstract this stuff anyway, like SDL/QT/GTK/FLTK/etc. There’s not much re-engineering effort for these developers. Of course there’s more work to do for those writing the low level toolkits, but then again it needs to be said that many of them already work.
https://doc.qt.io/qt-6/wayland-and-qt.html
https://docs.gtk.org/gtk4/wayland.html
https://fltk.gitlab.io/fltk/wayland-devel.html
I’m not here to be an evangelist and say it’s perfect, but I do think you are exaggerating the problems for most users and developers. Not for nothing but I was highly critical of wayland for a long time…but I can admit that its making progress in the areas that I care about and most people won’t have a reason not to use it even if it doesn’t satisfy you.
I was a proponent of Wayland in lets say first decade. Then the realization hit things are not as great as promised. At that point i didn’t know what exactly went wrong but over years i realized it’s managed rather poorly and there was a lack of any sort of obligation involved, to make the transition easier. Like we will implement protocols, slowly and will take our time, you do the rest. Then there were hackathons for GNOME, on where some important features got implemented, next week KDE did the same amount of work again. Xfce for example didn’t even start to implement such features yet. If this is such great approach, why aren’t we doing it elsewhere too, for example in regards to audio. As for toolkits helping here, not much. Some libraries did emerge, but ended not being used as much. No wonder it will take 20 years of work to get to some semi workable fragmented mess. This is just not acceptable as a concept. It’s idiotic. Even if it ends up working in general it will still be a pile of never ending and fragmented mess, like a display server black hole sucking in sparse resources, over and over and over again.
In retrospective what in my opinion should have happened instead is X11/Xorg should have been replaced with something like Wayland/Worg. In this 20 years of development Wayland/Worg on where developers would be doing much of the heavy lifting and in 2024 we would have had a strong GNU/Linux display server/compositor stack that might have even be used in Android. Most of the DEs already ported to support it. For projects on where it makes sense to develop a custom Wayland compositors, for only them to do that.
Geck,
I understand your frustration with the approach, but this is quite customary for the way things get developed on linux. As much as we could complain about it, I don’t really see it changing…. FOSS gives different groups a lot of freedom to do what they want regardless of what anyone else wants, and that’s what ends up happening. Personally I still like this better than proprietary locked down software, but there are certainly pros and cons.
Ok, but this exact description could just as easily apply to X itself. That’s the thing, wayland, for all it’s faults is replacing a display manager that’s notoriously bloated and inefficient. And while it’s true that extensions have enabled some of this to be bypassed, it’s still a messy antiquated architecture and I think it’s due for replacement. I understand you may not like wayland, but for better or worse it’s the future while X is the past. Even if you have a better idea for wayland, you’re much too late to change the current trajectory.
Such generalization is not OK. It’s good to have a variety of software targeting some area, variety of desktop environment, variety of web browsers, programming languages …. This shows the ecosystem is healthy and thriving and here choice is good. Having a gazillion Wayland compositors, now that is moronic. The same as packaging the same application for gazillion times. This two areas are holding GNU/Linux adoption back and here we need better solutions. As for criticizing Xorg. Big, archaic, obsolete … fine. Still nobody felt the solution is to reimplement X11 for gazillion times. Here Wayland still has much to learn, before we can say GNU/Linux has a good display stack. Currently it’s a laughable fragmented mess and i am not talking about Xorg. Xorg can’t be blamed for wrong turns Wayland took, that is on Wayland to resolve. By i guess learning a thing or two from Xorg, on where Xorg did get it right.
Geck,
You’re welcome to feel this way, but fragmentation is to be expected with linux and FOSS. Like I said there are pros and cons but at the end of the way this comes with the territory.
X also has faults that shouldn’t be above criticism. You can feel whatever way you like about it. Whatever opinions you have about X and wayland are fine when it comes to deciding what is best for yourself…But regardless of your decision you need to know that it’s not going to stop more people and distros from choosing wayland as the future.
There are many valid criticisms of how Wayland has been managed as a project. However, the time to predict that it will never get there has passed.
Today, not in the future but today, the two leading desktop environments (GNOME and KDE) are both Wayland by default. The most popular desktop distributions are now shipping with Wayland by default and are considering removing support for X11 (other than Xwayland). The most influential corporate distribution (RHEL) just entered beta with its default desktop Wayland only as well. Distros like Chimera Linux that have not even entered beta are Wayland first by design. The most exciting new desktop environment (COSMIC) is Wayland.
In 2025, the majority of Linux users will be using Wayland and will not even be thinking about it.
Clearly, any DE that wants to survive needs a Wayland story. And we have seen progress on this from Cinnamon, MATE, LXQt, and XFCE. Support has not been that difficult for most of them as both GTK and Qt have supported Wayland for a while. When we exit 2025, only the most niche window managers (I guess we have to call them compositors now) and desktop environments will be X11 only. As a percentage, pretty much everybody will be running Wayland. Even other POSIX systems like FreeBSD and Haiku are on the train now.
It does not really matter what we think of Wayland anymore. It is not just the future. Wayland is the de facto standard display server technology in the Open Source world today.
Whenever I use a full DE, vs a WM, XFCE is my favorite and I see this as great news. It is the GTK desktop that most respects my preferences, and is just comfortable for me to use.
Wayland is the way forward as X has more latency and is inherently more insecure than Wayland. There is a reason people have moved away from attempting to improve X and money is not the primary reason. It’s code base is huge and cumbersome for new developers to come in and get up to speed with. That alone makes Xorg an impossibility for the future.
SeanEParsons,
I like XFCE a lot because…knock on wood…they’ve been much better than other desktops when it comes to not trying to reinvent the desktop…. Gnome are guilty of this and to a lessor extent KDE. XFCE just provides a good basic desktop that doesn’t shove new UI paradigms just for change’s sake. Many user hate this and I think linux users and windows users (mac users…?) have this in common. Don’t force us to change workflow and muscle memory if there’s no good reason to, we don’t appreciate that.
Considering the amount of users GNOME has vs XFCE, I’d say many (most?) of us do appreciate it.
@Anthill4991,
I will agree that there are apparently many people that do like GNOME, but I do think that GNOME’s numbers are bolstered because most of the major distros package GNOME as their default desktop. The main reasons they do that is because of GNOME’s predictable release cycle and momentum. While they all offer re-spins with other DEs there is a perception that they do not spend as much time polishing them. A lot of us on OSNews, including Thom, will install whatever DE/WM we prefer and modify it till it meets our workflow needs. I just appreciate how little work I need to do to make XFCE work how I prefer.
> A lot of us on OSNews, including Thom, will install whatever DE/WM we prefer and modify it till it meets our workflow needs
I use Sway, not GNOME. But your post is exactly missing the point – a lot of us on OSNews are so detached from all of those people who aren’t on OSNews. Trying to seriously draw a conclusion on which DE the average user appreciates based the general vibes in OSNews?! Next thing you know and Thom writes a piece about how Haiku is soon to be taking over all governmental offices.
This detachment could have been a cute thing that defines our niche, but it’s sad and dangerous because it often leads serious people to discard serious work done by serious developers for audience that just isn’t them, and it just contributes to how fractured the FOSS world is.
Anthill4991,
Popularity isn’t always about merit, sometimes it’s the network affect that drives popularity in a positive feedback loop. KDE wasn’t even a first class citizen in popular distros like Fedora until very recently.
https://fedoraproject.org/wiki/Interviews/KDE4
Clearly this has given gnome a marketing advantage and better support. Alternatives often have to exist against this backdrop. Still, despite this there actually were a lot of defections spawning desktop environments for users who didn’t appreciate gnome’s changes. Everyone’s free to have an opinion though and I can respect that.
Congratulations Xfce team! Looks like another solid release with tons of improvements and fixes. Add a simple and familiar interface and no surprise Xfce works so well.
As for Wayland support – I tried it myself. It works but there were some noticeable paper-cuts here and there. Not sure if they are coming from Wayland, labwc or Xfce, probably all of them. Importantly, there were no benefits from running Xfce on Wayland (or even Wayland itself), so anyone looking for the best Xfce desktop experience is better served by the X11 stack.
A couple years ago, I upgraded my Dad to Endeavor Linux(yes Arch – but I do the upgrades and it’s blissfully simple) and Xfce, combining two computers, one that was running a horribly old version of Ubuntu and GNOME, and one with a similarly crap slow Windows 10. It’s still going strong. I’m incredibly impressed how easy XFCE makes things while maintaining speed on older hardware. Also XFCE never seems to change things that don’t need fixing.
KDE and GNOME and Windows are simply too heavy for the hardware, and have a distracting emphasis on flashy attempts at being novel and “easy to use” that frankly are just confusing mostly for someone who thought Windows 95 was about perfect.