Fedora to stop shipping Xorg in Fedora 41.

[athenian200]

So, it looks like the first step to deprecating Xorg has been taken.

https://linuxiac.com/fedora-41-drops-th ... -sessions/

For anyone that uses Fedora, does this change affect you negatively? Are you able to work around it, etc? Just curious how bad things are getting over on the Linux side with regards to Wayland pushing Xorg out.

[Moonchild]

Of course it'd be Red Hat that does this first. I'm not surprised.

[Pentium4User]

To clarify, Fedora will continue to support Xorg in GNOME; it just won’t be included by default in new installations.

[mrnhmath]

Why would anyone continue be a guinea pig for Red Hat in the good year of 2024 is beyond imagination.

[athenian200]

Why would anyone continue be a guinea pig for Red Hat in the good year of 2024 is beyond imagination.

Oh, tell me about it. Honestly, they're doing a lot of stuff in RHEL 10 that is going to make things a bit harder for us...

[Pentium4User]

Can you give more details, please?

[mr tribute]

The big problem with Wayland is that developers tied security to a display protocol. Two things that have nothing in common. Why not do like Android and use native Linux features (each app runs as a separate user, hence isolated from other apps) to implement security regardless of display protocol chosen? Running X.org with user privileges (instead of root) is also a quick fix to many security problems.

[athenian200]

The big problem with Wayland is that developers tied security to a display protocol. Two things that have nothing in common. Why not do like Android and use native Linux features (each app runs as a separate user, hence isolated from other apps) to implement security regardless of display protocol chosen? Running X.org with user privileges (instead of root) is also a quick fix to many security problems.

That's exactly what I was thinking. The biggest complaint I always hear about Xorg is that it can capture user input globally... so okay, then if a distro cares that much about security, then don't let X11 see the keyboard, and encourage applications and window managers to handle accepting user input outside of X11 through some other library that captures input and is tied to the specific application. What, you say that would break some X11 features and piss users off? Well... what do you think Wayland is doing? You say a lot of them would just turn Xorg's input handling back on to fix compatibility issues? Well... again, why do you think they keep falling back to Xorg?

Not really sure Wayland did anything nearly as revolutionary as they say, and it seems like a lot of the problems are being fixed by slowly adding back features that Xorg already had while trying to improve on them... regardless, I worry it's a moot point and that Linux will move on to Wayland only sooner rather than later, simply due to inertia, Xorg deprecation, and expectations rather than because Wayland is actually better.

[Moonchild]

Wait... Wayland handles security? That makes absolutely NO sense. What a monumentally stupid idea.

[Pentium4User]

Not really sure Wayland did anything nearly as revolutionary as they say, and it seems like a lot of the problems are being fixed by slowly adding back features that Xorg already had while trying to improve on them... regardless, I worry it's a moot point and that Linux will move on to Wayland only sooner rather than later, simply due to inertia, Xorg deprecation, and expectations rather than because Wayland is actually better.

I am sure this will happen in near future. If they mayor desktops support Wayland, the userbase for X11 is small.

[vannilla]

Wait... Wayland handles security? That makes absolutely NO sense. What a monumentally stupid idea.

I haven't been in Wayland territory in a while so I don't know how things have changed, but a few years ago one of the first thing you could read on the official website to justify creating Wayland itself was that applications were not isolated regarding user input.
So in a way, yes, Wayland was designed from the ground up to make "secure" applications.

[Pentium4User]

Wait... Wayland handles security? That makes absolutely NO sense. What a monumentally stupid idea.

I haven't been in Wayland territory in a while so I don't know how things have changed, but a few years ago one of the first thing you could read on the official website to justify creating Wayland itself was that applications were not isolated regarding user input.
So in a way, yes, Wayland was designed from the ground up to make "secure" applications.

Which isn't something bad in my opinion. Think about a web browser that has a vulnerability that isn't able to read the input of an ssh password in xterm.

[Moonchild]

one of the first thing you could read on the official website to justify creating Wayland itself was that applications were not isolated regarding user input.
So in a way, yes, Wayland was designed from the ground up to make "secure" applications.

Look, I understand wanting to improve that, but that should not be coupled with rendering - they are architecturally very different animals and no matter how many smart people worked on it, that dichotomy will never be smart to work with and I remain of the opinion that doing so is a monumental mistake...

[Nuck-TH]

At this point "security" sounds more like excuse than anything else.

TBH at this point i don't see any merit in wayland except than it being "trendy" and "new".
Performance advantage is long gone ever since actual functionality began to be implemented. Everything needs significant rewrites to work with it slowing down or outright halting fixes or feature development.
But best of all - is that after more than 15 years of development it is still feature incomplete and only borderline usable, while Xorg continues to just work(c) and has much more functionality despite neglect...
Here is some good insight.

It is very sad to see such counter-productive developments, but at this point i have no doubts that it will be pushed through to all distros like it was with pulseaudio and systemd...

[vannilla]

Which isn't something bad in my opinion. Think about a web browser that has a vulnerability that isn't able to read the input of an ssh password in xterm.

I understand where you come from, but the point is that removing legitimate features from a desktop rendering protocol because of some hand-waving "security" based on hypotetical and very targeted examples is not good practice.

[Moonchild]

Here is some good insight.

Interesting read. I particularly find the following bits very relatable:

To Wayland's credit, they took the good parts of xcb and made it easier to use ... but you'll quickly run into the opinion of the developers. Xorg is essentially "mechanism not policy" ... Wayland is exactly the opposite. It's "policy not mechanism"... guides you to manage your windows in the "right" way. ... There's a ton of stuff you can do client-side on Xorg that you straight-up can't do on Wayland and probably will never be able to do because of the development philosophy.

For a concrete example, ... it's not uncommon for users to want to start an mpv window on a different monitor or something like that. The answers I got back from Wayland developers were not very encouraging and one of them even thought it was a mistake to allow fullscreen to go to a specific output! I ended up abandoning that since I wasn't in the mood to try and write an essay justifying how people should use their computers.

and...

You must design your render loop in a certain way. If you don't, prepare to face the pain.

Seems to me Wayland devs are trying to do things too "neat and clean", i.e. they are being too purist.
Having to justify use cases the protocol devs did not envision and are reluctant to accept is not providing a stable foundation for system-wide use "in the wild". Application developers don't want to be held back by a draconian "policy" born from purism.

So I see the big picture issue with Wayland being two-fold: they want too tight control and too clean use cases, and at the same time they want Wayland to govern a lot more than it effectively should.

[Pentium4User]

Which isn't something bad in my opinion. Think about a web browser that has a vulnerability that isn't able to read the input of an ssh password in xterm.

I understand where you come from, but the point is that removing legitimate features from a desktop rendering protocol because of some hand-waving "security" based on hypotetical and very targeted examples is not good practice.

I don't use Wayland and I don't agree with its design because it is much, much too restrictive and breaks everything we know.

[athenian200]

I don't use Wayland and I don't agree with its design because it is much, much too restrictive and breaks everything we know.

That is also my view on Wayland. It's just too opinionated a protocol. Don't get me wrong, I see the problems with Xorg... and honestly, if Wayland were about replacing that with something sane that more closely resembles the display stack on Windows or Mac and thereby makes things more portable, I would be all for it.

But it doesn't. It's so restrictive that applications being ported from Windows or Mac can't do all the things they can do on other platforms and are forced to limit themselves to work within the vision of Wayland's developers. I get the feeling that what's going to happen in the end is Xorg is going to be killed off, Wayland will be adopted, it will have so many problems that people will complain about it constantly, and then after that we're going to get another rewrite of the display stack that won't be compatible with either Xorg or Wayland.

My prediction is based on what happened with how Linux went from ALSA, to PulseAudio, and now it may be doing PipeWire. If there can be a PipeWare after PulseAudio, why wouldn't there be another display stack rewrite after Wayland? In the end, I wonder if all applications will wind up falling back to working within the constraints of various things that simulate Xorg, like XWayland, XMir, and Xwhateverisnext, and eventually the underlying display stack is abstracted away while all applications just target a fake Xorg rather than whatever is really running on the system... sort of like how all these various terminal applications all try to simulate a VT100 even though the underlying system is very far removed from being a VT100...

I feel like Linux as it exists in the mainstream is just not friendly to small projects and is designed around teams that can keep up with major platform shifts every few years.

[Moonchild]

In the end, I wonder if all applications will wind up falling back to working within the constraints of various things that simulate Xorg, like XWayland, XMir, and Xwhateverisnext, and eventually the underlying display stack is abstracted away while all applications just target a fake Xorg rather than whatever is really running on the system...

My question about that is: in what way would that be better than using Xorg directly? Adding a translation and abstraction layer is going to cause performance loss and bugs. Why would you want that in favor of using Xorg?

[athenian200]

My question about that is: in what way would that be better than using Xorg directly? Adding a translation and abstraction layer is going to cause performance loss and bugs. Why would you want that in favor of using Xorg?

Well, you wouldn't want that, but if the original Xorg is abandoned, Wayland doesn't work out well, and all the Linux graphics stacks wind up simulating Xorg badly, and thus a half-baked Xorg simulation becomes the common denominator between them all, guess what everyone is going to target?

If no one is willing to go back to the original Xorg because of security concerns, that leaves running something like an entire X server inside a sandbox for every individual application with performance loss and bugs because people are too afraid of running Xorg on bare metal and the alternatives are all insufficient.

I mean, I hate to say it, but it's a lot like how Mozilla basically gave up on security and stability in the browser, and just relies on sandboxing that has each tab run in a container so that when one goes down, it won't bring down the whole browser. This would be the OS equivalent of that if it happens on Linux.