The no future of X.org, FreeBSD becomes headless?



RHEL10 is dropping X.org.

Fedora 40 (to be released in 5 months) is already deprecating X.org for Gnome and KDE sessions. Fedora is an upstream project for RedHat. If RH drops something, Fedora is going to do that much earlier.

What should I read outside a ton of ill-placed sarcasm in your message?



RedHat leads Linux and it seems like Unix as well. They are the driving force behind and provide almost all the money, resources and backing for: X.org (which they are abandoning in less than 2 years), systemd, pulseaudio/pipewire, Gnome (and a multitude of related libraries), Wayland and a ton of other projects. They are investing in the Linux kernel a lot.

If RedHat decides something other Linux projects follow. There have been no precedents to the contrary outside of Devuan, a Debian fork with no systemd.

Basically Linux is RedHat. Without the company Linux would have been nowhere near where it's now (that includes Steam Deck which wouldn't be possible otherwise).



It's weird "it's not so bad" yet *BSDs and UNIXes have made very few contributions to Xorg over the past decade:


It's basically been a RH show.
I'm sorry if you took what I said as sarcasm. I didn't mean it to be sarcastic. I simply don't agree that your concerns are valid and think they are more akin to fud than a necessary discussion regarding FreeBSD. I completely disagree with your assessment regarding Red Hat and Linux as well. Again, this seems very off topic for the FreeBSD community forums.
 

With Wayland every desktop environment and window manager must create their own separate graphics server.

It sounds crazy and almost lame, but the Linux community has rejected the idea of a single Wayland server akin to the X.org server.

Now we have over 15 competing implementations and some libraries (wlroots, libweston and louvre) which strive to be the backbone of all Wayland servers, unfortunately that's not been happening. Gnome Mutter and KDE KWin don't use these libraries.

Considering how complex the protocol is, minor projects have been struggling (XFCE) or outright rejected (IceWM, JWM) to adopt Wayland.
 
I must say that in my experience Gnome/Wayland performance destroys the performance on X. But there are so many WMs/DEs and DMs that don't support it that are excellent. I currently use LXQT with OpenBox on X and find it to be super. :D
 
I'm sorry if you took what I said as sarcasm. I didn't mean it to be sarcastic. I simply don't agree that your concerns are valid and think they are more akin to fud than a necessary discussion regarding FreeBSD. I completely disagree with your assessment regarding Red Hat and Linux as well. Again, this seems very off topic for the FreeBSD community forums.

OK, please close this discussion if the issue is not a concern for FreeBSD. Considering all the comments here to the contrary it surely looks like you don't quite understand the gravity of the situation but I'm an alien here, and when in Rome, do as the Romans do. Sorry for bothering you.
 
If you'll read the article that was posted it says Fedora does not have plans to drop X11. That's what we were working off of here. :D And if you read from Red Hat's website you will see that Fedora is upstream. It's basically the testing for Red Hat. It's been that way since Fedora Core was launched.



https://www.redhat.com/en/topics/linux/fedora-vs-red-hat-enterprise-linux
Uhhh... Fedora is a derivative of RedHat.... Yet, the web sites do say that it's 'upstream', even Wikipedia...
explains how come:

After Fedora was forked from RedHat in 2003, the former was promoted to be the testbed for the latter. The former takes Open Source stuff and integrates it, while the latter takes that work, does QA on it - and takes it closed source...

Yeah, before reading that explanation, the 'upstream' part did not make much sense to me. Usually, when a software is forked, it's considered 'downstream'... Patches and other improvements are sometimes sent 'upstream' - something that Red Hat seems to be picky about doing these days.
 
Uhhh... Fedora is a derivative of RedHat.... Yet, the web sites do say that it's 'upstream', even Wikipedia...
explains how come:

After Fedora was forked from RedHat in 2003, the former was promoted to be the testbed for the latter. The former takes Open Source stuff and integrates it, while the latter takes that work, does QA on it - and takes it closed source...

Yeah, before reading that explanation, the 'upstream' part did not make much sense to me. Usually, when a software is forked, it's considered 'downstream'... Patches and other improvements are sometimes sent 'upstream' - something that Red Hat seems to be picky about doing these days.
Yeah, it's a weird circular relationship. I don't think it's as easily explained as a downstream or upstream since the projects are so closely developed. Also the language of downstream and upstream are used when referencing systems as well as software implementations and it gets a bit confusing.
 
OK, please close this discussion if the issue is not a concern for FreeBSD. Considering all the comments here to the contrary it surely looks like you don't quite understand the gravity of the situation but I'm an alien here, and when in Rome, do as the Romans do. Sorry for bothering you.
I understand the concern for Linux. This is like a systemd type of situation. The power of the mighty RH forcing it's methodology onto other systems. But these things tend to get taken way out proportion even spilling over into FreeBSD and I find that to be surprising.

It's also interesting that no matter how many bold or business oriented moves that big Linux companies make there is so often an explosion of debate when there are also so many forks. This part is not unique to Linux but happens all the time with all types of software. Trinity desktop forked from KDE 3.5 comes to mind. There are so many forks of software and so many difference methods of deploying different solutions on different system. I just don't understand the need to react with such alarm.
 
OK, please close this discussion if the issue is not a concern for FreeBSD. Considering all the comments here to the contrary it surely looks like you don't quite understand the gravity of the situation but I'm an alien here, and when in Rome, do as the Romans do. Sorry for bothering you.
I imagine that FreeBSD will probably offer both Wayland and Xorg, and get Wayland ironed out so that it's on equal footing, and a viable option on a FreeBSD install. Until that happens, might as well keep Xorg around as a usable alternative/backup option.
 
Considering all the comments here to the contrary it surely looks like you don't quite understand the gravity of the situation
So RedHat stops "supporting" X (whatever that means. Not adding updates, not doing whatever?) Then just like every other open source project it will fork (OpenBSD Xenocara as an example), others will step up.

And it was a pretty silly to start talking about systemd as if it were a good thing.
 
The Linux Foundation took up Xorg as now the main or secondary fork, which is still under the same permissive license, as FreeDesktop.org. I believe they still maintain that, as they maintain PulseAudio, Pipewire, Canberra, Avahi and plenty of LGPL licensed programs. Those aren't going away from the Linux Foundation. That's separate than what RedHat does.

Xorg is also still there as from the original repository from the foundation it came from, even if it lost progress for updates.

OpenBSD has Xenocara. I believe there needs to be a BSD Foundation (as differentiated from simply FreeBSD Foundation) for maintaining forks, homegrown and original software that's for all BSD's and any other OS which wants to use it. It would serve a purpose like the Linux Foundation, except be for BSD.
 
It's weird "it's not so bad" yet *BSDs and UNIXes have made very few contributions to Xorg over the past decade:
Matthieu Herrb (OpenBSD) was on the Xorg board as was Alan Coopersmith (Solaris). They have made many contributions and still do.

BSD made an entire fork of Xorg called Xenocara which hosts *many* local modifications to the Xorg project.

Just the Xorg-specific hardware abstraction layer alone I see quite a few updates (don't believe all the trash you read on the Phoronix gaming forums or reddit. Just look yourself)
https://github.com/openbsd/xenocara/commits/master/xserver/hw

A BSD lead Xenocara is a very likely future for UNIX/BSD. I think Linux should be a little worried about losing access to this.

The big question is, what "contributions" do you need? The guys in the UNIX world maintaining this stuff don't give a crap that you can't play your Steam DRM platform games at 200fps on your ultra-high res consumer gaming monitor or Apple landfill device. Everyone else has been quite happy with Xorg for the last decade and will continue to be for the next decade. This truly is what maintenance mode is.
 
RedHat announced two days ago they are abandoning X.org starting 2025.

What are you guys going to do?

1) keep using FreeBSD;

2) when not, use something stable that suits your needs.

'abandoning x-org' is going on for years, just like 'systemd is coming'. So what? There are plenty of OS and distro's to use proven or even ancient technology.

Development of computer techniques always comes with alternatives and a lot of time and alternatives to choose. Besides: FreeBSD is aimed at stability, not on promising techniques per se.

No Worries.
 
So RedHat stops "supporting" X (whatever that means. Not adding updates, not doing whatever?) Then just like every other open source project it will fork (OpenBSD Xenocara as an example), others will step up.

And it was a pretty silly to start talking about systemd as if it were a good thing.

I think that was directed at me "And it was a pretty silly to start talking about systemd as if it were a good thing." My bad, not my intention. I am not really a fan of systemd. I think it may have a place on server but not on workstation and I've only found in annoying in the past when I've used it.
 
OK, please close this discussion if the issue is not a concern for FreeBSD.

It's of interest, but not concern.

FreeBSD is not a Linux 'distro' and isn't dependent on what any particular Linux crew does.

Considering all the comments here to the contrary it surely looks like you don't quite understand the gravity of the situation

Why do you think it's grave for FreeBSD? It may be for some Linux variants, but we ain't one.

but I'm an alien here, and when in Rome, do as the Romans do. Sorry for bothering you.

No need for sarcasm. Thanks for the concern but don't panic!
 
I think that was directed at me "And it was a pretty silly to start talking about systemd as if it were a good thing." My bad, not my intention. I am not really a fan of systemd. I think it may have a place on server but not on workstation and I've only found in annoying in the past when I've used it.
Edit:
Nope, back up in post #23 OP talked about it.
 
A bigger concern is Mozilla dropping X11 support for Wayland. Despite the backlash against Wayland; more and more major open source vendors are moving towards it. Sure your old crusty Window Manager will still work, but we will be left behind with regards to what software we can use.

I think we should take Weston and steward development of it towards a reference compositor like Xorg IMO. First party support for FreeBSD is already there, and It'll be easier to add in our own FreeBSD specific improvements too. Why go backwards?

Sure we can do a fork and fix of X.org; but who wants to maintain that cesspool? X.org is really, really bad. Wayland is a much cleaner protocol, and better suited for modern input devices; despite the ridiculous fragmentation currently.
 
What are you guys going to do? Your graphical display server is now declared abandonware.
I'll continue running win11 as my desktop and lazily hacking hyperv/vmware workstation drm drivers so that I could finally see what wayland is about :D
 
Weston is the older and original compositor for Wayland. And I read that it's a mess. This might be the only compositor compatible with both libx and libxcb through Xwayland. I think Weston would be a stepping stone and good testing ground, similar to what Beastie7 wrote. Weston is the only compositor which can be tested for Xwayland for both libxcb and libx applications right now.

While, it's a good thought about Weston, I don't think that's the longterm answer, because I read that it's not well organized and not built as well as WLroots. WLroots got to learn lessons from Weston. So, whatever compositor will be, it needs to learn lessons from both of these compositors.

WLroots might only be compatible with libxcb and not libx through Xwayland. There's very few programs in libxcb, and some have parts or dependencies still in libx. So, WLroots could be tested right now for pure libxcb progams, through Xwayland. For reference, WLC is defunct which was replaced by WLroots.

I based possible compatibility with libx or libxcb from looking at WLroots and Weston's documentation for what it had dependencies for or what it had parts of for use with Xwayland being libx and/or libxcb.

Wayland can work for everyone. It's modular for use. I see it for using an updated Xwayland or other compatibility layer for libx and libxcb built applications. Why not have the better Wayland protocol replace X, so existing programs can run on that as is. It already shares the same drivers from Xorg. Not all compositors are compatible with Xwayland, libx or libxcb.

The compositor and components would be for a compatibility layer, so existing X11 (both libx and libxcb) applications which don't have a graphical toolkit can run natively on Wayland.

As long as the vast catalog of programs is largely under libx, we'll need Xwayland or an implementation for libx for immediate use. Hopefully it transfers to libxcb over time, so libx can eventually become defunct.

We should adopt a BSD compositor, and make a better Xwayland layer, along with making it compatible with libxcb and libx. So that everything works. In the meantime, it can wean off of libx to libxcb, for likely a period of 10 years. The whole time, let those run alongside programs which use the toolkits of qt, gtk and SDL. There's no reason to not have more toolkits.

I think we should take Weston and steward development of it towards a reference compositor like Xorg IMO. First party support for FreeBSD is already there, and It'll be easier to add in our own FreeBSD specific improvements too.
I agree, except it shouldn't be Weston, because it's the primitive version of a compositor. Weston needs to be that for testing now, but it needs to be a better compositor. Perhaps a cleaner one based on both Weston and WLroots.

In short, Weston is a current testbed for that, rather than the long-term compositor which may not exist yet to steward. If Weston were to be stewarded, it would be great to be stewarded only as a testbed, and getting X11 applications to run on it might be possible in less than a month on it. Weston can remain as a long term testbed.

Wayland doesn't have to be looked at so negatively, because the compositor can be whatever type of compatibility layer or whatever else you want it to be, while being able to use the improvements of Wayland. Whichever compositors chosen, stewarded, or used, can leave out un-useful overly complicated Linuxisms, as they would be modular that way.
 
I agree, except it shouldn't be Weston, because it's the primitive version of a compositor. Weston needs to be that for testing now, but it needs to be a better compositor. Perhaps a cleaner one based on both Weston and WLroots.

Weston has the mindshare and branding for it being reference Wayland compositor. That'll go along way with trying to incite others to adopt it. It's also upstream, so whatever enhancements or bug fixes we provide, that wider open source community can benefit; much like OpenSSH, OpenZFS, etc. For the thousands of window managers and small team DEs out there we'll need a way to curate adoption too; call it WestonPorts or something. That'll give us an edge for possible FreeBSD migration too.

The rest I agree with.

Also I want to point out that Wayland has been used in other obscure use cases like kiosks and IVIs. Bringing FreeBSD to the forefront of Wayland development for more use cases is a win-win in my book. X.org is just becoming technical debt at this point.
 
Correct me if I'm wrong, but REL is unlikely to be used on the desktop in the first place, no? Isn't it mostly a server and enterprise used OS?
Correct.

We manage about 1400 RHEL boxes at $JOB. 1350 are VMware VMs and the few that are physicals are part of an Oracle ghetto. None have X.

Oracle is installed through an X installer. Some Oracle apps require an X server, so DBAs display to Xvfb. RH has removed support for utilities that Oracle has documented DBAs use to verify the environment before installing Oracle.

Oracle, through OEL (Oracle Unbreakable Linux) will probably support X for a long time because their DBMS installer needs it. I don't see X going away anytime soon.

I doubt RH wil continue to support even Wayland. Their customers are big enterprises and their focus over the last few years has been cloud through OpenShift and to a lesser extent podman.
 
I imagine that FreeBSD will probably offer both Wayland and Xorg, and get Wayland ironed out so that it's on equal footing, and a viable option on a FreeBSD install. Until that happens, might as well keep Xorg around as a usable alternative/backup option.
FreeBSD already has wayland in ports.
 
No fvwm2 no deal.

But you can always run Xnest on Wayland I suppose...
Agreed. I use CDE most of the time and maintain cde, open-motif, and fvwm2.

x11-servers/xwayland is the X server under Wayland. It's the Xnest of the Wayland world. I'm taking off for a couple of weeks over Christmas to spend some time with my mom when she turns 100 in a couple of weeks. I'll probably spend a few quiet hours playing with Wayland on my laptop. Maybe get CDE working under Xwayland. Just a matter of disabling xdm and starting seatd. A person can always go back by disabling seatd and re-enabling xdm. It's probably not much of a big deal. It'll certainly be a learning experience.
 
We need to consider the Gnome is not the primary desktop for much of the world. Remember, gnome3 veering off to produce the monstrosity it is today resulted in mint, cinnamon, and the mate desktops. GTK5 will probably result in similar forks and divergence.

There's a lot of software out there that is still reliant on gtk4, gtk3, and even gtk2. It will be quite some time before we see any of gtk5 or gtk6 (when that is even a thing) in the current pipeline. Browsers are probably the main concern as almost everything else can be replaced by software which relies on other toolkits. For the short to medium term, I'm not concerned. For the long term five to seven years, may ten, it's hard to tell. It's certainly worth keeping an eye on.

Then again, with the state of the world as it is, this'll probably be a minor annoyance in comparison other problems we'll be facing. Looking at the big picture it's not worth worrying about now.
 
Back
Top