When Time Comes for GTK5, It Might Be Wayland-Only on Linux
phoronix.com
phoronix.com
Breaking backward compatibility ought to be considered the cardinal sin of open source/free software, and they get away with it too much.
Yeah, compatibility is nice, but not when it means you have security flaws innate to the system, bloat and stuff you never want but have to keep around, etc... Sometimes the positives outweigh the negatives.
Wayland breaks a huge amount of apps, with no good porting story.
Most of us just want our apps to be able to keep working. Linux is not a place for a wizard in the high castle to disconnectedly architect the future of Linux GUIs.
But yes, I can absolutely say "Y'all are wasting a ton of time with this Wayland way y'all are doing things, you should go fix X instead."
While it's true no one can stop you telling people they should be working on what you want and not what they find valuable, you sound like a petulant child demanding people do things for you.
What I'm doing is pretty similar to dealing with, e.g. a land developer putting up a big shopping center near my quiet neighborhood. I don't have much legal recourse here, but someone who speaks against it isn't childish.
I believe a free/open source desktop OS that's easy for "regular" people to use is VERY important. One of the "selling" points is "stuff doesn't break, it will work the same way it does today as it will tomorrow." This could improve the user experience for a lot of people, this could get more people interested in tech, this could save a lot of e-waste. Again, this isn't mere ideas for me -- I do this with lots of people all the time.
Now sure, you could get on your tech high horse and say "well, all those newbs should catch up." Yes, I think the people who focus too strongly on comparatively small things like fractional scaling aren't seeing the bigger picture. And that is pathetic.
Not breaking stuff is a great goal, but it is a trade off, there is a cost to supporting the old, and it isn't always worth paying. Just sounds like your priorities don't line up with the people actually paying and doing the work. Your attempts to convince that it's a better option by telling them they are stupid actually are misguided if you think that'll actually work.
It's been what, a decade? And X still works fine. All I hear is this pathetic "fractional scaling" and "screen tearing!!"
For example, I’ve never managed to get it to be tear-free, on various hardware and software combinations over time. It got infuriating.
It’s part of why I switched to macOS.
Those are all functions macOS and even Windows have had for years (ever reloaded your graphics driver without shutting down applications? Windows can do that since Vista...) Since X does not have the right architecture for those kind of features it's unfit for modern desktops. Also, lots of X functions are no longer necessary or even practical to use anymore. Why is Windows remote desktop useable with an order less bandwidth than most X clients. (of course that's all the clients fault and not the X server, but that doesn't matter for users does it?)
If you want a vintage Unix workstation lookalike running 4 xterms and xeyes on a static desktop PC with a single ethernet connection X will do fine. But why force other people to look at the world in such a limited way?
I mean, most of those things aren't solved problems on Wayland either. The only thing we "solved" was screen tearing issues (fixed in most x11 compositors anyways...), configurable DPI is still extremely buggy, extended color spaces haven't even started development afaik and Xorg's input handling is still architecturally fairly sane. The 'problem' is that these hacks we used to use on x11 still feel better to use than their Wayland alternatives. Plus, once you consider all of the functionality, software compatibility and desktop environments we sacrificed for this meager victory, it seems utterly pyrrhic. Nothing of value was gained for the overwhelming majority of the software ecosystem.
> Since X does not have the right architecture for those kind of features it's unfit for modern desktops.
I agree. Consider the following: Wayland doesn't feature support for AppIndicators, which both MacOS and Windows have had for... 30 years. Is that befitting of a modern window server? Nobody's asking them to support it in GNOME, but rather to offer simple hooks in the code for the vast majority (eg. KDE, i3wm, Sway, Budgie, GNOME 3, MATE, LXDE, XFCE, etc.) of desktops that would want that functionality? It's another anti-cooperative move on the behalf of GNOME's development team, which is why it's no wonder that the rest of the community has abandoned the prospect of working with them. x11 needs to die urgently, and watching Wayland fail to replace Xorg, one of most ill-concieved software systems ever designed, makes me want to pull my hair out.
> If you want a vintage Unix workstation lookalike running 4 xterms and xeyes on a static desktop PC with a single ethernet connection X will do fine. But why force other people to look at the world in such a limited way?
Be pragmatic. If GNOME went Wayland-only tomorrow, 90% of the Nvidia systems would break overnight, that's not just ostracizing "Unix workstation lookalike" users. The reason Xorg is still around is obvious: Wayland simply isn't complete. The vast majority of desktop environments don't support it. The vast majority of desktop applications run worse on it. I despise x11 for being as crusty and stiff as it is, but I can run it on all of my systems without any problems. I can't do that with Wayland, and until I can, I have no reason to switch. x11 is superior for my workflow, full stop. There's no appeal to use it when only one or two of my computers actually run it well, and the rest have horrible showstopping bugs.
Those have started development. They're not done yet. Compare to X11 where no one is working on fixing those things.
>Wayland doesn't feature support for AppIndicators
App indicators are not a wayland or X thing, the API has been in dbus for some 10 odd years now.
>but rather to offer simple hooks in the code for the vast majority (eg. KDE, i3wm, Sway, Budgie, GNOME 3, MATE, LXDE, XFCE, etc.) of desktops that would want that functionality? It's another anti-cooperative move on the behalf of GNOME's development team, which is why it's no wonder that the rest of the community has abandoned the prospect of working with them.
I have absolutely no idea what you're trying to say here, all those other desktops do already have their own way they implemented app indicators. For GNOME it is available via an extension. There is nothing anti-cooperative about this and nothing was abandoned.
>Wayland simply isn't complete. The vast majority of desktop applications run worse on it. I despise x11 for being as crusty and stiff as it is, but I can run it on all of my systems without any problems. I can't do that with Wayland, and until I can, I have no reason to switch. x11 is superior for my workflow, full stop. There's no appeal to use it when only one or two of my computers actually run it well, and the rest have horrible showstopping bugs.
I don't think this matters much, most of them are very small niche projects in the already small Linux niche. If you found bugs then report them, there is no other way they get fixed. Staying on x11 forever is not a viable option. Wayland probably won't ever be complete, but x11 is both incomplete and abandoned.
> Staying on x11 forever is not a viable option.
"Staying on GTK2 forever is not a viable option."
"Staying on GTK3 forever is not a viable option."
"Using native packaging forever is not a viable option."
"Using Linux kernel v4.19 forever is not a viable option."
I wonder how many times I've heard this mantra repeated before...
Right now, native packaging isn't a viable option for a lot of things and it hasn't been for some time, have a look at the success of Docker.
Wayland is not that. People are welcome to defend it all they want, but this futuristic technology won't replace everyone's setup. It's not inevitable. They made that choice when they dropped support for hardware that x11 handles without issue. My point is that people have used outdated technology for more petty reasons than just "it doesn't work on my hardware" before. If Wayland wants to be minimal and petty about their featureset, fine. They'll have a minimal and petty community to go along with it. Nothing is inevitable on the Linux desktop though, I'd hope most people on HN of all places would know that by now.
I am not speaking of you personally, this is in terms of distros as a whole. It has definitely not failed there, they will drop these packages when they become stale and it becomes too burdensome to support them for little benefit. Like, I don't know any common non-niche distros that are still shipping a 2.6 kernel. You can attempt to fight them but it will probably be futile, few have any reason to maintain that.
>Wayland is not that. People are welcome to defend it all they want, but this futuristic technology won't replace everyone's setup. It's not inevitable.
>Nothing is inevitable on the Linux desktop though, I'd hope most people on HN of all places would know that by now.
Not true, death is inevitable. What is really inevitable is that x11 will fade away. Everything dies eventually. The question is what will replace it when that happens, if not Wayland then something else that probably has worse hardware support. We went through this already with the drama around Mir. Or maybe the Linux desktop will just die because it was always obscure, unpopular and fragmented with a severe lack of developers.
>My point is that people have used outdated technology for more petty reasons than just "it doesn't work on my hardware" before. If Wayland wants to be minimal and petty about their featureset, fine. They'll have a minimal and petty community to go along with it.
I don't see why this point is significant to anything. Using outdated technology for similarly petty reasons is just furthering the same problem and putting yourself in the same petty communities.
I certainly believe that this is more likely than everyone transitioning to Wayland. And, when it inevitably does, I don't think the majority of people are going to pick Wayland to run their legacy systems on. If both desktop servers stopped development today, I reckon they'd have to pick the most-finished option.
X being 40 years old has nothing to do with when it's reasonable to break compatibility. The amount of time that a suitable replacement has been around does matter; Typical times are 2-5 years from a suitable replacement being available to deprecation, which is to say that we'll be able to deprecate X in 2040 when someone's built something that actually solves user needs, unlike Wayland.
They shall fix the flaws not create new ones. Why do i need GTK 1, 2, 3 and 4 on my system ?
Considering how many people have abandoned the GNU/Linux desktop in favor of MacOS, I'm not convinced "everyone" tolerates it.
But it's also unclear to me how you would even quantify this. What does being intolerant in this context look like, other than just not using gnome? Do you expect gnome users to picket and/or burn down the homes of the developers?
It's kind of a ridiculous statement, and I feel like it's demonstrative of the kind of asshole entitlement FOSS users often exhibit instead of say, getting involved to influence the direction of upstream development.
Abandoning GNU/Linux in favor of macOS for political reasons, specifically as a reaction to Gnome, makes no sense to me. The things that frustrate people about Gnome (low configurability, highly opinionated design, low interoperability or deliberate incompatibility with other ecosystems) are all things that are very much present in Apple software. Even other parts of this stack, which it is insinuated emanate from the same overbearing Red Hat influence (systemd, PulseAudio), were inspired by macOS on a technical level.
I understand macOS has other strengths, but idk why anyone would see it as a refuge from the perceived problems with Gnome, systemd, Wayland, etc.
So what I want to say is that it's subjective, don't say it's worse like some objective fact, when there are many people who don't share your opinion, just like there are many people that never liked Gnome (not now, not before), and they're all correct in that it's mostly about what works for each individual and that's it.
It depends?
It's not like there is any apps for linux desktop. Who cares if you break backward compatibility.
Otherwise, it's cool and all, but it's not a complete spec by itself since it leaves stuff like screen recording to implementations.
When distros are confident enough to say their implementation is ready, I'll accept it, till then, it doesn't seem to have any real big new features besides the secure isolation.
Edit: I do OBS screen recording.. didnt know it was such a big deal. Fedora 34,35, 36 ish ? has it working, but I imagine that is pipewire + obs. Its pretty easy.
It seems like it would be the target of a lot of disapproval, though.
Wayland is the go-forward display stack for Linux. Get with the program or get left behind.
Corporate users. There's zero incentive to cut functionality out of a commercial library, so they'll probably gladly support Xorg until their customers stop paying for it.
> extant issues of Wayland clients and compositors have no bearing with whether GTK5 should drop the X11 backend; file bugs for toolkits, compositors, and the protocols...
Ah yes, that does seem like a sane way to address concerns with dropping Xorg entirely. Even the Titanic shipped with lifeboats, they'd be wise to have a backup in case their "unsinkable" window server breaks on many people's configurations.