Gtk 5 Might Drop X.11 Support, Says GNOME Dev (theregister.com) 145
One of the GNOME developers has suggested that the next major release of Gtk could drop support for the X window system. The Register reports: Emmanuele Bassi opened a discussion last week on the GNOME project's Gitlab instance that asked whether the developers could drop X.11 support in the next release of Gtk. At this point, it is only a suggestion, but if it gets traction, this could significantly accelerate the move to the Wayland display server and the end of X.11.
Don't panic: Gtk 5 is not imminent. Gtk is a well-established toolkit, originally designed for the GIMP bitmap editing program back in 1998. Gtk 4 arrived relatively recently, shortly before the release of GNOME 40 in 2021. GNOME 40 has new user-interface guidelines, and as a part of this, Gtk 4 builds GNOME's Adwaita theme into the toolkit by means of the new libadwaita library, which is breaking the appearance of some existing apps.
Also, to be fair, as we recently covered, the X window system is very old now and isn't seeing major changes, although new releases of parts of it do still happen. This discussion is almost certain to get wildly contentious, and the thread on Gitlab has been closed to further comments for now. If this idea gains traction, one likely outcome might well be a fork of Gtk, just as happened when GNOME 3 came out. [...] A lot of the features of the current version, X.11, are no longer used or relevant to most users. Even so, X.12 is barely even in the planning stages yet.
Don't panic: Gtk 5 is not imminent. Gtk is a well-established toolkit, originally designed for the GIMP bitmap editing program back in 1998. Gtk 4 arrived relatively recently, shortly before the release of GNOME 40 in 2021. GNOME 40 has new user-interface guidelines, and as a part of this, Gtk 4 builds GNOME's Adwaita theme into the toolkit by means of the new libadwaita library, which is breaking the appearance of some existing apps.
Also, to be fair, as we recently covered, the X window system is very old now and isn't seeing major changes, although new releases of parts of it do still happen. This discussion is almost certain to get wildly contentious, and the thread on Gitlab has been closed to further comments for now. If this idea gains traction, one likely outcome might well be a fork of Gtk, just as happened when GNOME 3 came out. [...] A lot of the features of the current version, X.11, are no longer used or relevant to most users. Even so, X.12 is barely even in the planning stages yet.
X11 just works and is stable (Score:5, Insightful)
Re:X11 just works and is stable (Score:5, Insightful)
Depends on what you mean by "just works." No one is working on the X11 server and the client libraries (libxcb) other than maintenance. There will be no new version of X11, no new features. X11 has enormous amounts of cruft in it that have accumulated over the decades, and much of it has been bypassed. I read once that there are only three or four individuals on the planets who understand very much of X11's internals, and none of them are actively working on it at this time. And X11 has no security model, really, so any app can spy on pretty much all other apps. I'm surprised we haven't seen malware that sniffs passwords from browsers because that would be quite easy to do.
But I agree, I like what we've been able to do with X11 up until now and until there's widespread support for app remoting (likely at the toolkit level, in GTK, Qt, etc), I will continue to stick with X11. I would have thought that app remoting would have been a high priority on the roadmap a long time ago as that's something that many Linux power users rely on. Also I've always liked the fact that X11 had window managers. Don't get me started on client-side window decorations.
Re: X11 just works and is stable (Score:3)
I really dislike client said window decorations.
I much prefer all of my windows to have the same buttons in the same place.
Re: (Score:2)
I really dislike client said window decorations.
I much prefer all of my windows to have the same buttons in the same place.
You may have overlooked that this is not done by X11. It is done by the window manager. Depending on window manager you can configure whatever decorations you want and place them wherever you want.
Re: (Score:3)
Re: (Score:2)
X has the concept of untrusted apps which are nicely isolated from the others. In principle this also works fine. The problem is that modern extensions such as Xrender were never put into the set of extensions those untrusted are allowed to be used.
X really dies due to a lack of interest caused by some decision at Redhat to declare Wayland the future and spend all resources there, in combination with a lot of FUD which seemingly convinced everybody else to also not work on it anymore. But even of 1/10 of
Re:X11 just works and is stable (Score:5, Insightful)
I found no benefit to Wayland, while it remains a major step backward. GNOME/Wayland requires a user to be actively logged in on the console to enable remote desktops. This is a major regression all by itself. The display still tears moving windows around, but the key difference is that while X had all these weird ways to configure and reduce tearing, Wayland has none. Client side decoration really sucks, and not just for consistency. I sometimes get flickering decorators on other windows from other applications when the front window changes state; gtk client side bugs become windowing bugs.
When I first experienced X at the Berkeley sun lab, for me the dream was to be able to pull up desktop sessions and run real desktop application anywhere in the lab with a display for headless servers and devices thru xdm, This concept got replaced with an incredibly heavy crappy browser that has to on it's own locally re-implement all desktop app support features of an OS every time it launch, to constantly refresh content and run crappy ui code written in javascript that we now call web applications. This really sucks.
Re: (Score:2)
When I first experienced X at the Berkeley sun lab, for me the dream was to be able to pull up desktop sessions and run real desktop application anywhere in the lab with a display for headless servers and devices thru xdm
Back in the day, the existence of the X terminal was a massive advantage. While workstations cost literally thousands to tens of thousands, you could get an X terminal with a big screen for around a thousand bucks (maybe $1500 for a really nice one) and it would need no software maintenance whatsoever. For a while I used a Sun ELC netbooted from a 486 running Linux (xkernel ftw) as my only machine. The 486 was our household server, it replaced a former resident's NeXT Turbo Slab. I could remote netscape bac
Re: (Score:3)
X11 has enormous amounts of cruft in it that have accumulated over the decades
There's some pixel drawing functions that were considered big in 1987. There are not "enormous amounts of cruft". There's a bit of old stuff which hasn't been removed because it's still used a bit. The old stuff is fixed and static. It's not being updated and it's not really getting in the way.
Re: (Score:2)
I presume the cruft is backwards compatibility with things that can cause ongoing issues, particularly with the security model.
The X approach to seamless remoting royally sucks at higher latency links, working well only over latencies that you achieve within a single building or campus. It sucks less than the solutions for Wayland (which is currently a big question mark), but it's not good by any measure except that it does handle 'seamless' operation. In theory, in Wayland this could be a feature of a com
Re: (Score:2)
Waypipe is one solution to seamless Wayland app remoting. I suspect it will become the de facto standard.
Re: (Score:2)
At your comment I decided to try it to see how it performed. Unfortunately it just hung, and I can't really discern why so it doesn't bode well..
Re: (Score:2)
Well it is early days for waypipe. In the meantime, before GTK 5 drops X11 support, when using wayland you can transparently remote apps using the X11 fall back, which is automatic on KDE and Gnome, giving the best of both worlds (except for my focus follows mouse issues). At least with KDE and Gnome (and soon Mate), there is no discernible difference to the user between running on X11 or Wayland in all functional aspects, except that Wayland will be slightly faster and smoother and support some things be
Re: (Score:3)
But that X sucks over low latency links has nothing to with the "X model of remoting" as the X protocol is asynchronous. It is just that clients never exploited this which they could now with XCB. I am not sure Wayland will ever support good remoting as the meta information to do this well simply does not exist at the server.
Re:X11 just works and is stable (Score:5, Insightful)
It's in maintenance only exactly because it works and does what we need it to do. The same reason we don't see a lot of innovation in dinner plates and flatware. The last time someone tried to innovate flatware we ended up with the abomination known as the "spork". Why do we act as if going into maintenance is somehow waiting for death for software? It just means the project is complete.
Perhaps one day there will be a feature complete replacement that removes accumulated cruft that actually nobody uses, but we're not there. Wayland could have been that answer but instead of being feature complete, that project decided to gaslight us.
Re: (Score:2)
Does it though? Yes it works for you and for me quite well. But what about multiple monitors and high DPI on some monitors but not others. If you've used Windows or Mac in such configurations you'll know X11 is lacking to put it mildly. There are also the issues of hotplugging monitors and GPUs and switching between integrated and discrete GPUs which is hackish compared to Windows or Mac. Although I have no such hardware myself so maybe things have gotten better recently.
Yes perhaps one day Wayland will
Re: (Score:3)
I'm not sure why the Wayland devs have such a sense of entitlement. And Yes, gaslighting. First the claim was that nobody displays apps remotely, then that Wayland supports it just fine. When backed into a corner, well it WILL support it fine one day. Then back to well it supports it fine now, by which they mean you can runm XWayland to remote display apps that support X11.
Re: (Score:2)
Re: (Score:2)
Re: (Score:2)
I would have thought that app remoting would have been a high priority on the roadmap a long time ago as that's something that many Linux power users rely on.
That was and in fact remains a massive point of friction with Wayland. When it was announced there were literally no plans to provide any kind of network functionality, the party line was that VNC should be good enough for anyone. While it's possible to remote a single app with VNC, it's a hack which uses X11 to do the heavy lifting, so this represents a fundamental lack of functionality if the goal is to not use X.
Fundamentally the biggest problem with Wayland is that it does not add anything, it only take
Re: (Score:2)
The "malware" should be an X11 client that you allowed on your display.
While *technically* true, the 'allow' is implied by virtue of beginning execution in an context the user controls. Which basically any malware would be running in. Without sufficiently granular security controls to, say, get the user's consent for grabbing the keyboard versus the ability to interact normally with a display. Screen scraping, grabbing keyboard without focus, preventing screensaver from locking input, all are security issues in X.
Re: X11 just works and is stable (Score:2)
Any application running on the same host running in the user's process namespace can read the memory of any process in that space. It doesn't matter what your display server allows or not, you can just read and write directly to memory. None of this new Wayland crap actually stops malware.
Re: (Score:3)
This is incorrect. It used to be the case, but most modern distributions by default do not permit that anymore (except for root or with the appropriate capabilities being conferred).
Even if it were correct, you referenced namespaces, and those are increasingly employed to further provide isolation.
I will concede that this demands a more friendly mechanism to manage permissions, which isn't really seen outside the mobile operating systems. E.g. sometimes you want an application to be permitted to inject mo
Re: (Score:2)
No it can't. Processes are all isolated in their own address spaces. Only through specific inter-process communication channels (including explicitly shared memory) can processes interact with each other. That's kind of the whole point of how things work on Linux.
Re: (Score:2)
Standard Xorg on Debian does require authentication. Here's what happens if you try to start a program that can't read your .Xauthority file:
$ xeyes :0.0
No protocol specified
Error: Can't open display:
Re: (Score:2)
In theory GTK probably feels they have done that:
https://docs.gtk.org/gtk4/broa... [gtk.org]
I have yet to personally try it though.
Client-side rendering (was Re:X11 just works...) (Score:2)
Right now, everything is rendered in the client app and uploaded to the server as a pixmap
That's only done by brain-dead X11 clients. The X protocol itself supports rendering on the server no matter where the client is.
Re: (Score:2)
AFAIK, Android Auto and Apply Carplay are not X11 apps.
Re: (Score:2)
ssh -X
Re: (Score:2)
You can wish for higher-level reinvention all you want*, but at the end of the day, the "inefficient" pixel-moving Just Works and doesn't need reinventing all sorts of application to keep working.
Speaking as someone who recently tried to do remote X to a site 50 km away even with gigabit connectivity... it really doesn't "just work". X11 is poor at handling latency. Xpra does a better job, but open enough windows and the inefficiencies of just grabbing the output become painfully obvious again.
Re: (Score:2)
That's why I said 'poor at handling latency', and while you are right to say that it's better than nothing, the fact remains it's not so great that we can't come up with something better. I mentioned Xpra as it terminates the X-ness of everything remote and use an entirely different scheme to operate remotely. Latency isn't great for remote desktop ever, but X exacerbates it massively compared to other protocols. Of course Xpra may address the latency, but if you use remote applications heavily, it quick
Re: (Score:2)
I think you could implement Xpra completely at the client side and use only X to speak to the server. The X protocol is flexible enough and asynchronous. There is simply nothing in it which would prevent good performance even over low latency links. (In fact I started to implement this but could not work on it for a while because of more important things)
Re: (Score:2)
I can't claim to have the low level understanding of X to really deny the possibility, but I am skeptical given that advances haven't been made on this front except projects that opt not to use X to connect the remote application to a display (notably Xpra and NX).
Re: (Score:3)
It's not bandwidth, it's latency. The point is that while local networks can benefit well from X forwarding, high latency networks are now able to handle enough bandwidth to deliver serviceable remote application, but X assumptions drive latency sensitivity which cannot easily be addressed in WAN context.
Nowadays, it's worthwhile to have a solution that actually has a hope at remote operation, since it's now possible, though X can't really deliver on it.
Further, X also can't cope with a network interruptio
Re: (Score:2)
I bought the xlib reference and programming manuals sometime around 1999-2000 . That was enough to convince me that X was badly in need of replacement. I remember writing a little GUI toolkit in C just so that I could avoid dealing the the hell that is Xlib.
Re: (Score:2)
That was enough to convince me that X was badly in need of replacement.
Why? We're talking about the very low level access to the windowing system.
I remember writing a little GUI toolkit in C just so that I could avoid dealing the the hell that is Xlib.
Why on earth didn't you use a toolkit?
Re: (Score:2)
It doesn't just work. You can hate on Windows, but for my workstation that has a 4k monitor and 1980 monitor it scales just fine. X11 doesn't allow desktops to scale them worth a shit. Wayland is at least working towards that. I want to know if they've found ways for it to work well with NVIDIA graphics cards yet.
Re: (Score:2)
I want to know if they've found ways for it to work well with NVIDIA graphics cards yet.
The onus is on NVidia's side, and they're getting there. It works at this point, but there are problems- particularly in things using Xwayland.
Re: (Score:2)
Re: (Score:2)
It's stable enough to use as a desktop (I am), you may just run into problems firing up glxgears and such.
SSH+X (Score:2)
Re: (Score:2)
Re: (Score:2)
You do realize that XWayland is just X but it uses Wayland as it's display driver, right?
Re: (Score:2)
Why yes, he does. That's why he said if you need to run X11 apps on Wayland you can use Xwayland.
Re: (Score:2)
Unless some dumbass removes the X backend from the toolkit...
"Old" (Score:4, Insightful)
lso, to be fair, as we recently covered, the X window system is very old now and isn't seeing major changes
You know why...?
Re: (Score:2)
Generally because no one understands X11's internals. And it is old and a lot of cruft. Modern UI toolkits bypass nearly all of X11's features because they don't offer things like anti-aliased fonts or window compositing. All of this is down using extensions to X11 that bypass nearly all of it. In some respects using those hacks ends up with the worst of all worlds.
Re: (Score:3)
This comment is completely misleading. At its core, X is a very flexible and well designed remote pixel buffer management protocol. That some part like old line drawing API is no used anymore by modern apps and kept for backwards compatibility is not argument at all to throw it away. One could simply deprecate and later remove this part (or also keep it around for sake of compatibility). That X could be extended and evolved using Xrender etc to support modern clients is a strength and not weakness. Networ
Re: (Score:3)
Modern UI toolkits bypass nearly all of X11's features because they don't offer things like anti-aliased fonts or window compositing
X11 certainly offers window compositing. The font server doesn't do much for most people any more, but if the current codebase is even half way modular, then code to draw into pixel buffers should be very well separated from code to manage pixel buffers. So the font handling code should be mostly harmless.
All of this is down using extensions to X11 that bypass nearly all of it.
Re: (Score:2)
Higher than standard resolution screens, e.g. 4k monitors with 200+ DPI, are a problem with X. Just look at the Arch wiki page about it: https://wiki.archlinux.org/tit... [archlinux.org]
Every random framework and desktop has to support it separately, and it's pot luck if a particular app you use decides to scale well or not.
Re: (Score:2)
Higher than standard resolution screens, e.g. 4k monitors with 200+ DPI, are a problem with X. Just look at the Arch wiki page about it: https://wiki.archlinux.org/tit [archlinux.org]...
What I notice reading your citation is that most of the problems are with GNOME and GTK+. Maybe it's the GNOME team that has a problem with X. Or, you know, making software people want to use. They have been failing badly at that for some years now.
Re: (Score:2)
To be fair I haven't tried the latest versions, but I found KDE and Wayland to have issues with scaling too. Not as bad but not as good as Windows.
Mouse wheel setting is still my bug bear.
Re: (Score:2)
Higher than standard resolution screens, e.g. 4k monitors with 200+ DPI, are a problem with X.
Well, it's a problem with the toolkits for sure. X reports the DPI for each monitor via xrandr, the toolkits often don't make use of it. Though it sounds like they don't in Wayland either so it's not really a failing of X in the X vs Wayland debate.
Re: (Score:2)
I remember hacks to do anti-aliased fonts before Keith Packard came up with the composite extension, which itself is a bit of a hack, although one that has worked for many year now. And the hackish nature is the problem. Like I said nearly all of X11 is no longer used, yet is still present, and old problems remain.
Like you I know very little about X11 internals, and so my attempt to describe problems is obviously lacking. But there are a few people who know a lot more than I do. I recommend you check ou
Re: (Score:2)
composite extension, which itself is a bit of a hack,
What's hacky about it?
He explains very clearly the limitations of X11 and its architecture and how modern extensions basically bypass a lot of X11 internals to get stuff done.
X11 doesn't have internals: X11 is a protocol. The Xorg server, which is the main implementation of the protocol has internals. It sounds like they botched the architecture and there's a cleaner architecture in there waiting to get out, such that the core the extensions are built on
Re: (Score:2)
Besides that, if someone wants X11, they can use XWayland. The end user experience is hardly any different - set DISPLAY to something, run the app and it pop
Re: (Score:2)
That would require that GTK continues to support X11. XWayland is just an X server that displays on Wayland.
Why is it that everything Wayland touches turns into a combination of gaslighting and pulling the rug.Why is it always trying to pull a fast one?
Re: (Score:2)
Re: (Score:2)
You know why...?
Yeah, everyone knows why. The problems preventing it from developing further and meeting modern use cases are fundamental to the design. You can't put wings on a Ford Focus and call it a plane and expect it to fly, you need to design something from scratch.
X11 can meander in maintenance mode for all anyone cares. Just don't expect a modern feature rich window system.
So some X11 features aren't widely used... (Score:3)
... therefore we ought to toss it? Perl has features and language elements that a lot of people don't use. So does Python. Word processors and spreadsheets have a lot of features/elements that only a few people use. Heck, just about every language and application out there does. Let's toss 'em all!
(Also: "X.11"? I've been using X Windows for as long as I've been using UNIX and I've never seen it written that way. "X11"? Yes. "X.11"? Never.)
Re: (Score:2)
Worth mentioning that the space taken by those "unused features" is very small. X11 is not big.
Also worth mentioning that for truly unused features, the solution is to remove them, not rewrite the whole thing from scratch.
Re: (Score:2)
Worth mentioning that the space taken by those "unused features" is very small. X11 is not big.
No one is talking about space. The "features" being discussed here are fundamental to the core design. People looked at removing them and decided it would be easier to write a completely new window manager from scratch, including actual X11 developers who moved onto other projects as a result.
Re:So some X11 features aren't widely used... (Score:5, Funny)
X11 is like a roof that was originally a 40-foot timber yacht. You've turned it upside down and fixed it to the tops of the walls. Gradually, over the years, you've patched in the holes until it only leaks when the wind is in the West. You've figured out how to get a flue up through the thing. You've nailed a TV antenna to it, and sealed around the cable with silicone. When you put it there you never bothered to take the decks out, so it's almost impossible to get into and work on and the structural elements, optimised rather for the sea than for housing, make it not very useful for storage. You still have to repaint it with pretty expensive paint every five years or so, else it starts to rot, and for some reason it attracts lots of confused-looking seagulls.
Anyway, look at all the features! It's got a winged keel, a 200hp diesel engine, and a gorgeous timber and brass wheel. All the fittings are marine-grade stainless, the rigging was all almost brand-new when you installed the thing and in her day she'd do 27kt reaching across a good wind. Don't actually use much of that any more, of course, but still...
Technically the keel still violates local planning ordinance, and technically it still smells quite a bit of fish. But it's been there for 15 years and it works. There's no need to replace it.
What's that love? You want to build an extension? Ah.
Re: (Score:2)
You are not wrong; however, where is the real roof at? It is absurd to remove your "boat roof" if you do not actually have a roof to install.
Throw away bad things? Sure.
Leave yourself with nothing? Are you dumb?
There have been two serious attempts at replacing X so far... and they all seem to be insufficient in some way. Worse than X itself. So, what are we supposed to do if we want to continue using something other than Microsoft or Apple stuff?
Re: (Score:2)
... therefore we ought to toss it?
Nothing is being tossed. It's just not being supported in future development. In other news RollsRoyce is developing an aircraft engine and its only getting put on purpose built aircraft, not bolted to the underside of some wings that someone strapped to a car and expects to fly.
Keep using your old system. No one is stopping you. Keep using Perl, it's not going to stop other people from coding in Java.
Re: (Score:2)
and X doesn't have any kind of security model at all.
Well we've had the X security extension for a few years now.
Drama queens, going on 20 years now (Score:4, Insightful)
GNOME devs act like they're not happy unless people are hating on them. Just the other day they called the Arch people "clowns" for daring to document an env variable that let you use consistent file choosers, and they spitefully renamed it and locked the thread to stop discussion.
If you really want to make a GNOME dev mad, go in a bug report about a removed feature/config and ask them how often they break features they personally like. It's always features the think people are wrong for liking.
Re: (Score:3)
Gnome is not a platform worth building on. You are literally better off writing your own GUI library than using GNOME. (although there are better alternatives). Even Motif is better than Gnome.
Re: (Score:3)
GNOME is crap, yes. But it is not the GUI library: GTK is the GUI library.
GTK was really nice ... back in the 1.x and 2.x days, before the GNOME developers starting calling the shots: giving widgets weird behaviour and destroying theming.
And BTW, having coded both, I really prefer GTK+ 2.0 over Motif.
Re: (Score:2)
This does seem to be the case. They even note "it's only 10% of our users".
There's a reason I don't use gnome. About the only GTK thing I use regularly is GIMP.
Re: (Score:2)
Indeed. That is why I consistently ignore these morons. Sure, occasionally I will use a GTK application, but my window-manager is fvwm and not some bloatware with delusions of being a "desktop environment" or some such crap. If they decide that X11 is not good enough for them, then I will stop using their stuff. They have nothing especially worthwhile and I expect many projects will simply not move to GTK5.
If they do (Score:3)
They can kiss GNOME bye bye
Re: (Score:3)
I honestly didn't know GTK 4 was even out. Did they completely depreciate GTK 3 features this time?
"get wildly contentious"? (Score:4, Insightful)
No? This discussion isn't going to be "wildly contentious".
If Gnome wants to dump X11 (not X.11, there is no such thing as X.11), then they should feel free to do so, and then stop being a viable window manager for a good amount of people.
And old and no changes? Yes? Because it "just works"? X11 certainly has its problems, but as a display server used in millions of machines, those are all well known, and not really an issue.
This smells a lot like progress for progress' sake with not other reason than "PROGRESS NAO!".
No reaons to use GTK unless developing for Gnome (Score:3)
Seems to me unless your goal is to develop a Gnome application, there's not much reason to use GTK when building an application. So in many respects it doesn't matter what GTK devs do, as longs as Gnome functions well. If you want to develop a nice, modern GUI app, and one that can run on multiple platforms, Qt is the only game in town, really, and they will continue to support X11 (and many other backends) for quite a while yet I think.
But for those hating on Wayland, have you tried it recently? It works rather well. Heck even sshing to a remote box and running apps works... but that transparently falls back X11 for that functionality, so no more remoting at all with GTK 5.
My only problems with Wayland have to do with glitches that come about from doing things like focus follows mouse, something Wayland developers really don't want you to do. KWin on wayland and focus follows mouse is supper buggy. Hard to say whether it's a wayland problem, or a kwin problem.
Re:No reaons to use GTK unless developing for Gnom (Score:5, Funny)
But for those hating on Wayland, have you tried it recently? It works rather well. Heck even sshing to a remote box and running apps works... but that transparently falls back X11 for that functionality
In your paragraph where you tried to say what was good about Wayland, you only said things that are bad about Wayland.
Re: (Score:2)
But for those hating on Wayland, have you tried it recently? It works rather well.
So works about as well as X? Except when it doesn't:
Heck even sshing to a remote box and running apps works... but that transparently falls back X11 for that functionality, so no more remoting at all with GTK 5.
This seems like a colossal flaw:
My only problems with Wayland have to do with glitches that come about from doing things like focus follows mouse
Re: (Score:2)
Re: (Score:2)
But for those hating on Wayland, have you tried it recently? It works rather well. Heck even sshing to a remote box and running apps works... but that transparently falls back X11 for that functionality, so no more remoting at all with GTK 5.
That's the thing. The only thing Wayland brings to that table is people who now want to burn the table. It works because X11 is still there. All Wayland contributes is not getting in the way. Now the GNOME people want to get in the way in a big way by poisoning GTK, so I guess Gnome is going to have to go. Just as well, the first thing I do on a new workstation install is install just about anything other than Gnome to be the desktop in order to get some work done.
Re: (Score:2)
But for those hating on Wayland, have you tried it recently? It works rather well. Heck even sshing to a remote box and running apps works... but that transparently falls back X11 for that functionality, so no more remoting at all with GTK 5.
So it works, except for a major core feature of X11 that Wayland doesn't implement, and actually needs to use X11 to provide? This is exactly what we were complaining about when Wayland was announced, and they said it wouldn't have network transparency. If the argument for Wayland is that X11 doesn't do things, and that Wayland is newer and therefore somehow inherently more maintainable and secure (which is a fantasy to begin with) but then you need X11 for basic functionality anyway, so you're now actually
Re: (Score:2)
I recently tried Wayland. I didn't chose to but it happened automatically when I switched to Gnome. Everything seemed to work ok. Then I tried to share my screen in a Teams meeting. The option was gone. So I switched back to X. Everything works exactly like it did with Wayland. Except now I can share my screen.
Sure, that was probably only the result of a Teams bug. But what is a user to do? Decompile Teams and fix the bug?
I don't see how Wayland brings anything to the user except a questionable future for r
Re: (Score:2)
But for those hating on Wayland, have you tried it recently? It works rather well.
I've read it still doesn't work on FreeBSD.
Practically speaking, this is low impact (Score:2)
I'm going to assume that anyone who's deploying Linux to multiple workstations is using some sort of LTS release. So, practically speaking, those folks won't have to deal with this for at least a decade* (and that's i theyf're even running a graphical desktop at all).
Well, assuming IBM doesn't pull the rug out from under people again... and maybe that's not a good assumption to make, given recent history.
But let's face it - there aren't exactly a huge number of standalone Linux desktop users. There's a reas
Re: (Score:2)
For me, the year of the Linux desktop was ca. 1995.
So⦠(Score:2)
Re: (Score:2)
Everything.
Re: (Score:2)
Everything.
Probably not. More like "nothing that matters". Because things that matter will not make the move and stay on older GTK or move away from it.
Re: (Score:2)
What would break is that GNOME based on GTK5 wouldn't work on X11, only on Wayland. Meanwhile there's several things that X11 does that Wayland doesn't by design, and one of them in particular (remote network display) is actually implemented using X11 — and is something that the Wayland devs originally said we didn't need, then said Wayland would do, and then said Wayland already did even though it doesn't. It's also important to a lot of people who use it regularly.
I dropped support for GNOME long ago... (Score:2)
GTK already uses backends (Score:2)
So I wonder the reasons for getting rid of the backend. Is it a case that it is hard to maintain, or that it just won't work very well with something they hope to implement
Re: (Score:2)
Or to force everyone else to do it their way or no way at all. The whole Gnome and Wayland projects stink to the heavens of that.
Re: (Score:2)
Well you see, running Linux is all about choice, the question is who's.
BSDs (Score:2)
Well the, when/if this happens it is good bye Linux forever and off to a BSD. I have used and now using a BSD on a spare (as I type). Who remembers when GNOME took over gtk, people were saying something this would happen ? Gnome people said "no it won't", well here we are.
Compared to "modern" Linux, BSD is like opening a Window after a Thunderstorm cleared out all the humidity.
Seems Microsoft won, Linux Foundation have been chasing Microsoft like a dog chases a car. With all the recent changes, Linux is
Re: (Score:2)
Well the, when/if this happens it is good bye Linux forever and off to a BSD.
What? How does BSD help? Wayland is on BSD these days too, and GTK5 on BSD won't support X11 any more or less than it does on Linux. If this change occurs, and you want to use GTK5 apps on BSD, you will still need to use Wayland.
Compared to "modern" Linux, BSD is like opening a Window after a Thunderstorm cleared out all the humidity.
Haha no. Maybe if you're comparing to a stock redhate or ubuntu install. But Debian is plenty modern, and you can still install it without systemd, even without resorting to Devuan. (If you want to run GNOME for some reason, Devuan is still a good idea. But you don't need GNOME to r
Wither remote access? (Score:2)
I mean, X is weirdly complicated and bulky overall, but it works very nicely for running GUI apps remotely. Where's the robust and universal replacement?
I suspect ... (Score:2)
Re: (Score:2)
That makes zero sense. When commercial Unix software vendors are concerned about license compliance, they use a license server, typically flexlm. It doesn't matter where you run the program, it limits the number of copies which can run simultaneously. You could only reasonably believe this if you had never been a Unix sysadmin.
Gnome. (Score:2)
I can't wait to see the gnome x-11 replacement.
That'll make it easier to NOT HAVE GNOME ON MY SYSTEM.
Re: (Score:2)
If they are missing like in wayland for the most part I can already use Windows.
Wow, but then you'll miss all the amazing GTK 5 apps!
Re: (Score:2)
Those are the domain of the window manager, those haven't been a feature of X or Wayland.