XWayland Nukes The NVIDIA EGLStream Backend

[MorrisS.]

which is an argument to stick with xorg that, unlike wayland actually has support for VR with vulkan.

Afaik that has nothing to do with explicit sync, its just waylands flawed design.

Wayland has just introduced... Xorg has been running for 30 years now, but is going to be deprecated. When directX12 were introduced the major gaming titles were based on DrictX9. So? Newest technology needs its time to be integrated. Wayland is superior to Xorg bud needs time to be integrated adapting the system to it. Do you want to return 32bit just because time ago was the most widespread among the applications? IpV6 is less widespread against IPv4m so why not stay with IPv4? Because it is obsolete. As well as Xorg is obsolete. It works but it's obsolete.

[relsi1053]

Wayland has just introduced... Xorg has been running for 30 years now, but is going to be deprecated. When directX12 were introduced the major gaming titles were based on DrictX9. So? Newest technology needs its time to be integrated. Wayland is superior to Xorg bud needs time to be integrated adapting the system to it. Do you want to return 32bit just because time ago was the most widespread among the applications? IpV6 is less widespread against IPv4m so why not stay with IPv4? Because it is obsolete. As well as Xorg is obsolete. It works but it's obsolete.

wayland is old tech and has been around for over 15 years

[Jaxad0127]

Wayland by design doesnt even support VR.

You mean this? https://www.phoronix.com/news/Wayland-DRM-Lease-Merged

Several compositors support it.

[mSparks]

You mean this? https://www.phoronix.com/news/Wayland-DRM-Lease-Merged

Several compositors support it.

that is one aspect, but as pointed out above it was designed more than 15 years ago for single screen embedded devices, doesn't have any meaningful mechanism to extend it and hasn't changed much since then. "by design" it overs nothing for VR application development and in fact, for things like detecting if a headset is connected is designed to hide such details from applications.

several "vendors" have hacked sketchy support on outside of and despite wayland, but that still leaves it not supported and buggy more often than not - vs "just working" on X11.

plus, as we see from this thread, chances are anything developed for wayland would be removed after a few years anyway "because it got in the way".
No one will ever risk any kind of meaningful investment on that modus operandi.

But apparently all that is negated by the fact wayland supports explicit sync, just interested why and to who that actually matters.

[Daktyl198]

3 I probably dont undertand this statement. People (including Nvidia) tried to rewrite Gnome and KDE backends for EGLstreams and showed that it does not work without serious revisions to EGLstreams that never materialized.

4 i dont understand this either. You probably dont have much exposure to industry standards. They tend to be intentionally slow to ensure compatibility and buyin of all shareholders. The fact that updates are made means the standard is well and alive.
If you are looking for slow: look the time scales needed to update internet protocols ...
Also, examples of dead pojects that cannot be updated: VNC, X11, ...

3. My point here isn't to push EGLStreams or anything, merely to point out that the GBM vs EGLS debate had no overall effect on the timeline for Wayland adoption, as there were other blockers that wouldn't have been sped even if the debate never happened. Things like screengrabbing, copy/paste, etc.

4. Internet protocols are actually updated fairly fast. Adoption rate is an entirely separate matter. But scale compared to other projects was never my point. I was just saying that, objectively, the Wayland protocol extensions needed for day to day usage of a desktop have taken a very long time to come out mostly due to endless debates and bureaucracy. This, in turn, has been the #1 reason for delay in Wayland adoption. This was meant as a reply to the first comment, which implied that the GBM vs EGLS debate severely set back Wayland development/adoption, but it's just not true.

[mSparks]

mostly due to endless debates and bureaucracy. This, in turn, has been the #1 reason for delay in Wayland adoption.

It seems to me more "infighting" than that.

Wayland was 100% designed for lightweight embedded systems. It never saw widespread adoption in that market because mobile device manufacturers (google/ apple/samsung etc) were entirely capable of rolling their own, and did exactly that.

Then there is a seperate "faction" that want to utilise it on desktop, but don't actually have the power to add anything to the design a desktop actually requires, and even seemingly dont use or have any idea why people use linux desktops and servers in the first place.

This whole EGL debarcle is then just a manifestation of that infighting.

"X.Org development documentation glossary defines EGL as "Embedded-System Graphics Library".^[5]"

[Daktyl198]

It seems to me more "infighting" than that.

Wayland was 100% designed for lightweight embedded systems. It never saw widespread adoption in that market because mobile device manufacturers (google/ apple/samsung etc) were entirely capable of rolling their own, and did exactly that.

Then there is a seperate "faction" that want to utilise it on desktop, but don't actually have the power to add anything to the design a desktop actually requires, and even seemingly dont use or have any idea why people use linux desktops and servers in the first place.

This isn't really true. Wayland was developed by many X.org developers who got tired of maintaining an outdated specification that they didn't have the "authority" to update. It was always intended for desktop usage, but was modeled off of existing mobile/embedded display systems. The main problem with Wayland is that the people who designed it purposely made it as lean as possible ("here's how to draw a frame, and push it to hardware. Done") with the intention of bypassing the giant mess that X.org had become.

Unfortunately, those people did not even attempt to come up with possible solutions for the rest of the subsytems X.org was in charge of, which is quite a few. The main one being anything having to do with input. This left the rest of the ecosystem to try and hash out a standard together without any sort of "leader" in the space who could take charge, unlike what we had with X.org.

I wonder what a future with Mir would have looked like, as it set out to solve most of the issues that Wayland did regarding latency and round trips but without changing the "display server" dynamic that we had all become used to.

[anda_skoa]

Wayland was 100% designed for lightweight embedded systems

As Daktyl198 wrote it was designed in a way to would allow a light weight implementation for such systems, however even the base specifications already has, for example, support for multiple outputs which are much more common on desktop or workstation like setups.

It never saw widespread adoption in that market

It is actually widely used in the market as soon as a device needs to be able to run more than one fullscreen application. It has completely replaced X11 there.
UI system vendors in that area, e.g. the Qt company, even have Wayland compositor frameworks to make custom compositors easy to write.

because mobile device manufacturers (google/ apple/samsung etc) were entirely capable of rolling their own, and did exactly that.

These vendors are rarely involved in embedded projects due to their relatively strict control over the platform.
Embedded device vendors use a extremely wide range of hardware options and mostly "roll their own" Linux via Yocto or similar systems.

Then there is a seperate "faction" that want to utilise it on desktop, but don't actually have the power to add anything to the design a desktop actually requires

Actually that is the primary "faction".
The Linux core graphics developers (who used to work on Xorg) as well as the GNOME and KDE desktop shell developers.
Most protocols are directly derived from the needs the latter two faced in creating their respective desktop shell product and modifying their applications to run on them.

[mSparks]

This isn't really true. Wayland was developed by many X.org developers who got tired of maintaining an outdated specification

That is absolutely not true. original wayland developer is basically Kristian Høgsberg

Kristian Høgsberg (710 commits), Benjamin Franzke (32), Tim Wiederhake (22), Pekka Paalanen (20), Peter Hutterer (18), Callum Lowcay (17), and Tiago Vignatti

https://www.linux-magazine.com/Online/Features/What-s-Taking-Wayland-So-Long

"By contrast, Høgsberg summarizes the initial goals for Wayland by saying, “The core idea is that all windows are redirected, we can do all rendering client side and pass a buffer handle to the server, and the compositing manager runs in the display server."

X11 is maintained by Mesa, Intel AMD and Nvidia plus a few others
wayland span out of xorg development during the rise of mobile devices with limited or no network connection - where X11 made no sense. Its direct competitors are surface flinger on android and whatever ios uses. All the other original wayland developers have long since abandoned it afaict.

the direct equivelent to xwayland that has and will end up seeing widespread use is called matter, and is a colab between a large number of big industry players including google and samsung.

https://en.m.wikipedia.org/wiki/Matter_(standard)

[Daktyl198]

That is absolutely not true. original wayland developer is basically Kristian Høgsberg

Kristian Høgsberg (710 commits), Benjamin Franzke (32), Tim Wiederhake (22), Pekka Paalanen (20), Peter Hutterer (18), Callum Lowcay (17), and Tiago Vignatti

And he submitted it to the X.org Foundation for consideration and they adopted it and worked on it together with him. Yes, he did a lot of the original work, but it was very quickly taken up and iterated on by core X11 developers. The vast majority of the work in the last 5-10 years has been by X.org devs working in tandem with compositor developers.

"By contrast, Høgsberg summarizes the initial goals for Wayland by saying, “The core idea is that all windows are redirected, we can do all rendering client side and pass a buffer handle to the server, and the compositing manager runs in the display server."

Yes... and? This is almost exactly how rendering worked under X11 using window managers/compositors, except that the X.org server was a middleman. The client-side compositors did all of the rendering and just passed a completed frame to X11 for displaying. Then X11 would pass the frame around inside of it's 27 extensions before finally displaying it. Wayland simply removes the "passing the frame around" step. This is why the X.org foundation decided to pick it up and run with it (with some changes to the protocol).

X11 is maintained by Mesa, Intel AMD and Nvidia plus a few others

No, it's maintained by the X.org foundation who accepts patches from anybody, including Mesa, Intel, AMD, and Nvidia developers.

wayland span out of xorg development during the rise of mobile devices with limited or no network connection - where X11 made no sense. Its direct competitors are surface flinger on android and whatever ios uses. All the other original wayland developers have long since abandoned it afaict.

Again, once Wayland was adopted by the X.org foundation it was ALWAYS intended as a replacement for X11 on the desktop. Yes, it was also useful in embedded/mobile systems, but core Wayland developers were calling it the future for Desktop as well almost right away. Also, Wayland's only "direct competitor" is X.org/X11. And they're both maintained by the same group of people. Most of the original Wayland developers are still active in Wayland extension development, if you took the time to look at the git issues. And those wayland developers still maintain their respective systems on X11, they just don't actively develop on it anymore. It's in maintenance mode.

the direct equivelent to xwayland that has and will end up seeing widespread use is called matter, and is a colab between a large number of big industry players including google and samsung.

https://en.m.wikipedia.org/wiki/Matter_(standard)

Now you're literally just saying shit. Matter is a message passing standard for IoT devices that has literally nothing to do with display protocols at all. In fact, it's specifically aimed at devices without screens. It's not an equivalent to Wayland, X11, or xwayland in any way, shape or form.

With that last "point", I know you're just trolling. And if not... I seriously hope somebody teaches you reading comprehension and critical thinking skills.