Wayland

>systemd, it takes away the freedom to chose witch tools i can use for certain task
>wayland, it takes away the freedom to chose witch tools i can use for certain task
my argument is pretty much consistent desu

Okay list 3 tools you can't use on wayland and reasons why you can't port them over.

xrandr i would have to write a compositor, convince every DE and app dev to adopt it and ignore the main selling point of wayland (compartmentalized security) forcing me to develop a SELinux like module for a fucking protocol.
xdotool i would have to write a compositor, convince every DE and app dev to adopt it and ignore the main selling point of wayland (compartmentalized security) forcing me to develop a SELinux like module for a fucking protocol.
xbindkeys i would have to write a compositor, convince every DE and app dev to adopt it and ignore the main selling point of wayland (compartmentalized security) forcing me to develop a SELinux like module for a fucking protocol.

Does anyone have a reply to these? What's the answer to 's question, and does have a good reason to be worried?

I'm pretty sure X is never truly going away, though even in the shroter-term future more compositors will be developed along the lines of existing X window managers as Wayland picks up more support. It already has Gnome and KDE, as well as Sway being a drop-in replacement for i3. But as far as I'm aware, that's basically it at present and it's missing a functional replacement for *box WMs.

What do you mean?
Copy from Krita (which is part of KDE project) and insert it into gedit (just an example)? Or what is "KDE program" and "GNOME program" are then?

As a Fedora user, who has the option of both, I run on X right now simply because it seems much more responsive on my hardware. And this is running on foss drivers.

This. Wayland was released way too soon. It's still barely usable for fucks sake.

Copy paste from gedit into kate, copy paste from gimp into krita, copy paste from KDE's web browser into GNOME's web browser, with different data types (text, images)

Copy paste from software built on Qt into software built on GTK.

Does that work as well as it does in X?

when using kde/gnome on top of wayland, I think he means.

it wasn't supposed to take this long to replace it, was it?

There's a Wayland protocol, xdg-decoration for this already. KDE and Sway support it. GNOME does not because they have their heads up their asses (no surprise there). Just use KDE or Sway and you won't have this problem.

>xrandr
Your wayland compositor would need to provide a tool to provide this functionality. Other DEs/applications have nothing to do with this. It exists on all the decent Wayland compositors already.

>xbindkeys
An equivalent of this on Wayland makes no sense. There is no centralized server model, so there can be no """global""" way to set keybinds. Of course, you can just set keybinds with your compositor and there functionally won't be any difference.

>xdotool
This would be the hard one to actually replace. It would need a protocol and the applications would have to write support for it. As far as I know, no one has really done any work on this front, so you are out of luck here.

Well I tried with text (kate/gedit) and image (gimp/krita) and everything worked.

It's taking longer than I think most people would have liked, yes. It's understandable though. Wayland rips away a ton of things that Xorg does for you "for free" and goes for a very minimalist approach. To write a Wayland compositor, you have to implement nearly everything from the ground up which is a ton of work. That being said, Wayland desktops are far, far more usuable now than they were a couple of years ago so progress is being made.

Another setback is that Nvidia suffers from NIH and refuses to properly support Wayland. They do it in some hacky way that doesn't respect the proper APIs.

Just to add to this, the only difference I noticed between wayland and xorg (I tried switching between them for some time) is that wayland will tell you to fuck off if you try to run anything with GUI as root.

Yes.
The Wayland protocol does not have a central server like X11 has, and all the functionality has to be implemented directly by the compositor (=Wayland term for a program that does the same thing that an X server, window manager, and compositor would do in the X11 model).
There are libraries that reduce the imminent wheel-reinventing (e.g. wlroots), but that's just a band aid and doesn't solve the central problem.

Can you give actual arguments instead of just copy-pasting buzzwords you read on Jow Forumslinux?
That way we can determine if have any actual technical knowledge of X and Wayland, or are just an autistic retard.
My money is on autistic retard, by the way.

That's the subject of an ongoing feud between GNOME and literally everyone else. KDE proposed server-side decorations with a negotiation protocol that would allow the client and compositor to work out how to display things properly. GNOME is pushing for client-side decoration which are an unmitigated disaster - you can see that today even on X when running GNOME applications with a different window manager.

It wasn't supposed to be replaced. The beauty of X is that it's extensible. X11 was finalized 1987. Every new thing introduced since then has been implemented as an extension. And the system works, as evidenced by the fact that Wayland, after a decade of development and being hailed as the second coming of Jesus, is unable to surpass it in performance.

>wayland will tell you to fuck off if you try to run anything with GUI as root
This is intentional. Wayland only allows clients from the same user to use the display since the model is fundamentally different. X could elevate permissions with sudo (not recommended) or via polkit (which is a bit of a clusterfuck).

>It wasn't supposed to be replaced.
It is. Wayland is was written to replace Xorg. Xorg is meant to be on legacy support for a very long time though.

>And the system works, as evidenced by the fact that Wayland, after a decade of development and being hailed as the second coming of Jesus, is unable to surpass it in performance.
There are some people that oversell Wayland's hypothetical performances gains over Xorg. Wayland is a more efficient protocol, so technically it will perform better but the difference is not huge or dramatic. Benchmarks of Xorg outperforming Wayland are usually actually Xorg vs XWayland.

>It is. Wayland is was written to replace Xorg. Xorg is meant to be on legacy support for a very long time though.
That's borderline wrong.
For most of the time it was thought there was going to be X12, which would be backwards-compatible of course.
Wayland was actually just a hobby project of Kristian Hogsberg before the majority of Freedesktop latched onto it because it agreed with their vision of how the Linux desktop should work.
Then suddenly X.org became old. insecure, unmaintainable, and in swift need of being replaced. Which is ironic because X.org is maintained by the same people (Freedesktop) since 2004. So when they say the code is a clusterfuck, they really mean that they failed at their job, but pinkie promise they will do better next time. Like, for sure, dude.

>There are some people that oversell Wayland's hypothetical performances gains over Xorg. Wayland is a more efficient protocol, so technically it will perform better but the difference is not huge or dramatic.
No one cares for hypothetical, theoretical, or technical. "More efficient protocol" is a meaningless buzzword unless backed up by actual measurements. So far the benchmarks show Wayland is a rather pointless exercise in futility.

>That's borderline wrong.
Now you are just making up history and twisting it. Yes, there was originally going to be an X12. However it was abandoned because it was found that the changes need would be too dramatic and drastic. Instead, they opted to write something completely new, Wayland, which was meant to replace Xorg.

>So when they say the code is a clusterfuck, they really mean that they failed at their job, but pinkie promise they will do better next time.
Xorg is a clusterfuck because it's existed since the 80s and the way graphic rendering is down is dramatically different now than it was in the early 90s. The code became a clusterfuck because extensions were constantly added to Xorg, and lots of old, ancient legacy stuff was poorly maintained and rarely ever used. Fundamentally, Wayland's model of direct rendering is a more minimalist, simple approach compared to Xorg's big server. It is easier to maintain and more futureproof since it offloads a ton of the things Xorg did to compositors.

>No one cares for hypothetical, theoretical, or technical. "More efficient protocol" is a meaningless buzzword unless backed up by actual measurements. So far the benchmarks show Wayland is a rather pointless exercise in futility.
There's more to life than how well your games perform in Xwayland. You are more than welcome to stay on Xorg forever. Like I said earlier, it'll certainly be on legacy support for many, many years.

My first 2 and biggest problems with wayland:
- it hard-locks after a few dozen minutes
- it maxes out the number of x clients at 10
In other words, it's unusable garbage.

>consistent window decorations across all programs
That's because of gnome devs with their csd boner. Kde tries to push ssd, but gnome is more powerful.

>hard-locks after a few dozen minutes
>it maxes out the number of x clients at 10
No idea what you mean by either of these. The first one is certainly not true for me. As for the second one, there's no limit on the amount of clients that can connect to Xwayland either unless you mean something else.

>xmonad

Attached: feel.jpg (1280x842, 482K)

>Now you are just making up history and twisting it.
No. In fact I find the accusation particularly funny because...
>Yes, there was originally going to be an X12. However it was abandoned because it was found that the changes need would be too dramatic and drastic.
...right here you're trying to lie your way out of a tight spot.
X12 never left the "concept" stage because the extensions were working out nicely, and by the time it became relevant again most of Freedesktop was already on the Wayland bandwagon.
>Instead, they opted to write something completely new, Wayland, which was meant to replace Xorg.
Again, Wayland was a hobby project that was picked up because it fit a particular want. There were (still are) others at the time, like Arcan. Wayland is nothing special, really.
>The code became a clusterfuck because extensions were constantly added to Xorg, and lots of old, ancient legacy stuff was poorly maintained and rarely ever used.
This is actually a lot less of a problem than you're making it out to be. The old stuff hasn't been touched in years for the most part, and the new stuff does a pretty good job bypassing most of the old crap that you bitch about.
>Wayland's model of direct rendering is a more minimalist, simple approach compared to Xorg's big server.
That's an illusion. Most of the complexity is irreducible and the only thing Wayland did is make it the responsibility of the compositor.
>It is easier to maintain and more futureproof since it offloads a ton of the things Xorg did to compositors.
See? You admit it yourself. Wayland dumps the complexity on some other poor soul and people hail it as progress. And the fact that this causes even more fragmentation is conveniently overlooked.
>There's more to life than how well your games perform in Xwayland.
So tell me, what measurable quality favors Wayland over X? I'm interested in this.

Attached: fuckyou.gif (1680x1050, 642K)

>tfw there was better tech in 1990 but no one adopted it

Attached: NeWS.png (1000x500, 87K)

>X12 never left the "concept" stage because the extensions were working out nicely, and by the time it became relevant again most of Freedesktop was already on the Wayland bandwagon.
Are you retarded? You just admitted I was right. X12 never left the concept stage because what they wanted was too ambitious and required way too much change to be backwards compatible with Xorg. Hence, a brand new thing, wayland, was later created.

>Again, Wayland was a hobby project that was picked up because it fit a particular want. There were (still are) others at the time, like Arcan. Wayland is nothing special, really.
Why the fuck does it matter if it was a hobby project? You just described most of the free software you use. Linux was a hobby project to fit a particular want as well. Arcan was not publicly available for nearly as long as Wayland was. Yes, it technically existed, but nobody heard of it until much later.

>This is actually a lot less of a problem than you're making it out to be. The old stuff hasn't been touched in years for the most part, and the new stuff does a pretty good job bypassing most of the old crap that you bitch about.
It's a nonsensical model. Xorg is a bizarre middleman that everyone has to work around. Yes it "works," but not necessarily very well.

1/2

>That's an illusion. Most of the complexity is irreducible and the only thing Wayland did is make it the responsibility of the compositor.
>See? You admit it yourself. Wayland dumps the complexity on some other poor soul and people hail it as progress. And the fact that this causes even more fragmentation is conveniently overlooked.
This is the correct approach. The idea that the display server should be handling device inputs is 100% retarded and only people that can let go of Xorg think this is a good idea. Wayland, as a protocol, should only be concerned about rendering. And that is all it does and it is very good at it. Other things should be the responsibility of other programs. "Fragmentation" complaints are silly. Linux distros are already plenty "fragmented" and it's better for it.

>So tell me, what measurable quality favors Wayland over X? I'm interested in this.
Every frame is actually synced perfectly for starters.

2/2

If GNOME doesn't support it, does this also mean that anything built on GTK won't be able to use it either? That could be a hard problem to avoid since probably more than a third of all linux GUI software is built on GTK.

GNOME has an unfair amount of influence over desktop linux and they abuse the fuck out of it. I hope we'll be free of their bullshit some day

No, it's a compositor level thing. GTK applications have decorations on KDE just fine.

Yeh no xinput no use

Quick question, this probably belongs on /sqg/ (stupid questions general).

Does wayland depend on you being on systemd? Or can you use it with different init systems?

I dont even know software which explicitly rely on systemd init systems. I guess that answeres your question.

>Does wayland depend on you being on systemd?
No.

>Or can you use it with different init systems?
I'm using it with runit right now.

systemd ate a lot of things, so a lot of software unfortunately depends on the monstrosity. For example, logind is a part of systemd and it is a very common dependency (GNOME depends on it).

I agree with this, but there's no dwm for Wayland, so I can't switch.

lol at the mismatched themes, linux desktop still fragmented as fuck, as always

There's sway (i3 clone). That might be too much for you though.

Thanks anons for the quick reply. I've heard some desktop environments depending on systemd but haven't looked into since I don't use them. Unsure if it's just the way it's configured or if depends on features systemd has.

Well yeah, I'm using a GTK desktop and kate is a qt application. If I wanted to, I could install a qt theme and it would work well, but I don't really use qt apps so there's no point.
Windows has even less consistent theming

I tried it but I prefer dynamic tiling.

>logind is a part of systemd and it is a very common dependency (GNOME depends on it)
Didn't even knew how bad it's gotten. I just discounted logind and the likes as still being systemd but the fact that gnome depends on it is really cancerous. I went from early versions of ubuntu to systemd-less distro's and never even used it. I still have a cd with an old ubuntu release.

How is i3/sway not dynamic tiling?

oh god, i wish news and PostScript based desktops had caught on

i3 has manual tiling. You choose where the windows go, split horizontally or vertically, and make whatever layout you want.

With dynamic tiling you have a fixed layout and as you open windows they tile in that layout, you don't have to worry about where they go.

Ah, you can save layouts in i3 actually, but I don't think this works in Sway.

It's not about saving and restoring, it's about the window manager actually handling it on its own just as you open any new window.

>Windows has even less consistent theming
irrelevant

The benefit of it being a protocol rather than a full solution is that we will not be stuck with shitty ideas which seemed like a good plan 20 years ago (Xorg), the downside is that we will have many different implementations compositors running said protocol, at least until one emerges which is good enough to become a de facto standard.

The closest thing today is wlroots, it's a library upon which you can build your own compositor, with all the functionality you asked for implemented or being implemented.

github.com/swaywm/wlroots

Yes of course. I think its also part of the Wayland spec