>>52444786 You probably haven't been using it very long. It gets better ever year. More and more of the time you can just boot a distribution using Xorg and have it "just work". That certainly was not always the case. Every veteran Linux or BSD user can tell you about the hours of painful hacking at xorg.conf that was often required.
>>52444883 The switch from Nvidia's proprietary drivers to the open source AMD ones was certainly a vision. I went from having to manually add hacks to get the Nvidia driver to detect my monitor properly and even then having problems with some modes to everything just working with the open source one from the first boot.
Now if AMD could just get it up to speed and get at least OpenGL 4.3 support in place.
>>52445291 X11 is probably as far from >muh UNIX philosophy as no other widely used software.
X11 tries to be a flying pig with a kitchen sink and carries tons of obsolete cruft like the drawing functions that nobody uses anymore but every driver has to implement it. Network transparency is a farce. Every client draws itself anyway.
>>52450473 NO, WE SHOULD IMPROVE ON THINGS. IDIOT. WHY ARE YOU COMFORTABLE IN SHIT?
Wayland is clearly superior to X in every conceivable way, Wayland pretty much wrecks X. But no fucktards like you reject new technology because the current one still somewhat works, despite being a horrible fucking hack and slow.
>>52450473 Yeah, let's not try to improve on things because they already work... somehow. >who needs a chair when you can sit on a rock? >lol enjoy your proprietary clothes while I wear this rough fox pelt of +1 freedom >why even cook food? this raw elephant shit already tastes good enough
>>52450553 Just like on Linux, the issue isn't "if" it works but "when" it works. Currently Wayland isn't usable without a display server, which has to be provided by the WM. I'm pretty sure that so far only KDE and GNOME have actually created functional Wayland display servers yet. So you can use it, IF you're using KDE or GNOME.
>>52450501 >Wayland is clearly superior to X in every conceivable way, Wayland pretty much wrecks X. In gnome3 you can use wayland as backend but I had many problems with. I tried orbment about half year ago but back then some application were really fucking slow.
At least X works unlike any environment using wayland currently.
>>52451116 Well obviously a desktop environment designed for X is going to work like shit on Wayland. We have to wait until we have proper software for Wayland. Just because it's not backwards compatible with X, doesn't mean it's shit.
It's telling that while most of the stuff in the old Unix Haters Handbook is now irrelevant, the chapter on X11 is still largely true (although we no have toolkits like GTK+ & Qt and I haven't seen Motif in over 5 years+):
>>52452068 Gnome and KDE do, and those are the only ones that matter these days anyway >Also it's hard to find drivers that support wayland generally. Won't those be handled as a separate XWayland client?
>>52451952 so like every program I listed plus much more that you use? i'm asking why are you guys bitching about xorg specifically which isn't as bad as many other software you use daily and even call them good.
>>52452119 isn't it just experimental for those two though? >Won't those be handled as a separate XWayland client? not talking about X. Talking about drivers that will support the rendering pipeline Wayland uses.
>>52452586 >>52452649 >I don't see many posts like OP's about Sublime. That's because sublime is none of those things.
>X is just showing its age. X was badly designed from day one. http://www.art.net/~hopkins/Don/unix-haters/x-windows/disaster.html It's slowness is very measurable. There's hundreds of unnecessary calls made on every frame in X, i recall they can take upto 10ms to complete, easily causing dropped frames, or input lag. I really shouldn't have to explain why it is now a bloated shit show.
Being old maybe be the cause of some of this shittyness, but it's not a defence.
"Wayland is not really duplicating much work. Where possible, Wayland reuses existing drivers and infrastructure. One of the reasons this project is feasible at all is that Wayland reuses the DRI drivers, the kernel side GEM scheduler and kernel mode setting. Wayland doesn't have to compete with other projects for drivers and driver developers, it lives within the X.org, mesa and drm community and benefits from all the hardware enablement and driver development happening there."
so basically, if your X drivers are shitty, so are your wayland drivers.
>>52452807 People give out about Atom for being bloated, slow and buggy whenever it's mentioned. It's just that hardly anyone uses it.
>That section was written by someone who's mad that his own display server never caught on. You could be right anon. Maybe he owns stock in MS and Apple too. He's still right though.
>That sounds like someone using it wrong. ? That's how it works. Whenever a window renders something, this happens. It's one of the major issues with X, they developers know about this stuff, but because it's not functionally broken they never fix it for fear of breaking something. This is part of the reason for some members leaving to make Wayland. To leave all this useless cruft behind that has become so bloated and complex that no one knows how to fix anymore.
>>52452928 >Yes, but not with the same asspained tone. because no one uses it. It's like giving out about WinAmp or someshit. No one uses it so who cares? X is something people have to put up with a lot.
>Just all the events it causes? You can specify what events to capture in your X application. No, I'm talking calls that happen within X server itself. Doesn't matter if you capture them or not in your own application, X still makes them. It's just layers of crud that have been added over the years that no one can justify the time to optimise out, and if they could it would be a big job to do so with high confidence that you didn't break something. Again, the is why many X devs left to developer Wayland, because X does tonnes of shit it shouldn't be responsible for in the first place. And it's getting to a point now where it's so complicated that only about 3 or 4 people know how the whole thing works. It's not sustainable going forward.
>>52452986 >X is something people have to put up with a lot. For most people X's problem don't matter. A modern toolkit like gtk or qt just use directlrendering, moving most parts of X completely out of the picture. And the few mb a full xorg build eats is completely overshadowed by DE's bloat. What end-users struggle with are bad graphic drivers or shitty setups.
>>52452986 >No, I'm talking calls that happen within X server itself. Can you post your source? I want to know more details.
>>52453417 Tearing is a problem in most distros OOTB, and how you fix it depends on your GPU and WM. Oftentimes the fix comes with a trade off (input lag).
For example, I'm using an Nvidia card with Xfce. OOTB there's tearing, regardless of if I install Nouveau or proprietary drivers. If you want to get rid of the tearing 100%, you install a "3rd party" compositor like compton and configure it to use OpenGL. If you use OpenGL, then you get significant input lag because you have to wait for things to draw onto the screen because of the vsync. You can see it easily by watching your windows draw relative to your cursor (which moves instantly).
The other solution I've found is to force compositing via xorg.conf, which gives much less input lag but you end up with massive stuttering.
Part of it IS Nvidia's problem because the problem isn't as bad when I use IGP, but the issue is still there. And you can't really tell everyone "lol it's Nvidia's fault" when Windows and OSX don't seem to have an issue at all.
So no matter how you cut it, a real solution is needed badly in the Linux world and that's what Wayland+Weston are supposed to be.
>>52453494 >since the construction of the first computer. Show me an end-user complaining about lag in an application, that can demonstratively be traced back to inherent X flaws, and not the application using it wrong or using something bloated like gtk.
>https://youtu.be/RIctzAQOe44?t=1368 this links to ipc, not stuff the server does internally and also benchmarks a fucking gtk application, which does god knows what
>>52453938 Wayland is a protocol, it's not a compositor. Look at http://wayland.freedesktop.org/ if you actually care. But if you want, install weston (the reference compositor for wayland) and test it yourself right now. Windows are more responsive, there's no lag while dragging, there's no vsync issues. It's fairly clear that it's a superior desktop experience.
>>52453938 It's better because it isn't an incredibly slow bloated pile of shit. X's codebase is filled with hacks and fixes over the decades in an attempt to help it keep up with modern expectations of desktop UI. X is so complex now only about 3 people understand how the whole thing works, and there's tonnes of shit in there that can't be fixed at all now.
X was built for a different era. Now most applications really just need a framebuffer to render to, some information about the window and user input. That's it. All the networking transparent UI rendering stuff that X was built for is a thing of the past, and doesn't even work any more after all the new layers of shit added to it.
X causes very real input lag and rendering lag too as a result. Wayland takes all this away.
WM's don't need to rewrite everything from scratch either.
>>52454143 There's already a proposal on ML. Look for "Add Primary Selection Protocol Version" in this and previous month. It's not as straightforward because you don't want to let clients scan your primary clipboard.
>>52453546 I never noticed tearing problem with my amd card. my laptop with intel igp also doesn't have this issue.
> And you can't really tell everyone "lol it's Nvidia's fault" when Windows and OSX don't seem to have an issue at all. My amd gpu works 5 degrees cooler on windows, draws ~30 watt less power (although i only tested the overall system power draw, but what else could cause power draw?) Does that mean it's Xorg fault or amd's fault?
>>52446382 My dorm neighbor tried out pretty much every single option for running Wayland. Says sway is literally the most workable, even more than Weston. Still not ready and constantly breaks dialog windows and tooltips, but it's farther along than everyone, especially from independent compositors written mainly by one man.
>>52453938 With the X model before an application can do anything they have to pass their message to X. Then X has to observe the screen and decide if it can be put on the screen and if it can then it hands it off to the window manager who is actually responsible for placing it on the screen and the positioning of it and pretty much everything else. Then the WM sends the new information to X and X updates the entire screen and sends the information to the clients. The WM is basically the display server on the X model while the entire X server is just being used to pass messages. The problem is X really sucks at passing messages because it has a lot of functionality that cannot be removed and it constantly tries to use it where it should not be using it. Those functions introduce latency and cause graphical problems.
The wayland model removes the entirety of the X server and leaves behind a very minimal protocol that applications can communicate through instead. Application sends its stuff to the WM(now a full display server) and the WM just does it and then updates the screen and sends information back to the clients directly.
>>52454318 This is a bit misleading. X does very necessary things, but for old setups where direct rendering wasn't a thing and X had to have all rendering logic. When all your application does, is using OpenGL to draw everything using the GPU - then it does get in the way.
I can understand why a DEVELOPER could find it more "confortable" than the old, bloated, incoherent Xwindow.
But I totally fail to understand why on earth a USER should use it.
Wayland is not faster, is not stabler, has no network transparency, is incompatible with lot applications. It has no killer application, no "wow" feature, no compelling reason. And, last, but not least, we will probably have to keep both Wayland AND X for years, and years. Doubling the number of bugs and halving the number of developers.
All this, just for giving developers a new shiny environment to have fun with? Let's say it's ok, but why should I, the final user, care?
>>52455172 >Wayland is not faster, is not stabler Not true. It was the reason Wayland was created, too, because there is an improvement in both of these, or will be when it's smoothed out. >has no network transparency Neither does X. And it can be implemented in Wayland, it just doesn't have a default definition. >is incompatible with lot applications Which is why Xwayland exists. It solves half the problem, the other half with have to be solved by authors of compositors, admittedly. >It has no killer application, no "wow" feature, no compelling reason. It's overall better, distros will start switching to it eventually when it gets decent enough, the users won't care enough to switch back. There wouldn't be a reason to leave it either, do you see many people switching from ALSA to OSS nowadays? From systemd to other init systems? For most users, whatever comes by default, rules the day, unless it's blatantly unworkable. >And, last, but not least, we will probably have to keep both Wayland AND X for years, and years. Doubling the number of bugs and halving the number of developers. It's a better solution than staying on X for eternity. Also, GTK3 and QT5 programs are already supposed to be display server-independent, and most people use toolkits, there would probably be bugs, but most of these are supposed to be trivial. Old X-dependent programs will have to test their work both on X and Xwayland, sure, but it's not quite "every single program now has to spend twice the time on their bugs". Also, >>52455230 Developers really don't care about underlying X much, besides the fact what they are compatible with, and what not, plus some trivial stuff. It's mostly developed because it's supposed to be a faster, more stable, better thought-out way of doing things that will make every program work better. It isn't fully workable nowadays, but it's because of relatively little time spent on it. It will be ready one day, and then it will be objectively better than X, for users.
>>52455843 That's completely up to Nvidia and AMD. AMD is actually in the process adding wayland support to catalyst, they're making catalyst use the open source kernel driver which will bring wayland support as an unintentional but beneficial side effect.
Nvidia on the other hand is sticking to their guns and just rewriting their binary blob kernel module thing to add wayland support that way.
AMD will probably have wayland support on catalyst well before Nvidia adds wayland support to their driver.
>>52455955 Mir is not the same as wayland it's actually more like another X. It's an entire package, a display server/WM/compositor, support libraries, input handling, and whatever else is needed. Mir is designed to be "ready" at release. Wayland requires a bit more time to set up because although it has many libraries to help devs out it still requires other devs to make use of them.
>>52455981 I see you didn't read the rest of my post. Try reading it again, AMD and Nvidia are working on wayland support. Intel has had wayland support. As soon as wayland was released Intel supported it because their driver never relied on proprietary implementations of OpenGL.
Thread replies: 174 Thread images: 9
Thread DB ID: 425173
All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.
This is a 4chan archive - all of the shown content originated from that site. This means that 4Archive shows their content, archived. If you need information for a Poster - contact them.
If a post contains personal/copyrighted/illegal content, then use the post's [Report] link! If a post is not removed within 24h contact me at [email protected] with the post's information.