[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / asp / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / qst / sci / soc / sp / tg / toy / trv / tv / vp / wsg / wsr / x] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology

[Advertise on 4chan]


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: IsHuy.png (66 KB, 876x507)
66 KB
66 KB PNG
https://arewewaylandyet.com/
>>
>>79677467
i do not see unity on the list so i dont care
>>
>>79677518
Same, but with BSPWM.
Not gonna touch it until my preferred WM is reimplemented.
>>
when any keyboard/mouse sharing software works
>>
>>79677467
I'll switch once XFCE is ported
>>
>>79677534
honestly i am strongly considering learning gtk just to maintain unity.
can barely get it touch with the current devs.
>>
>doesn't work on 80% of desktop GPUs

This shit will never be relevant unless freetards stop crying about >le evil nvidia :(((((((
>>
>>79677620
you dont get it do? the code is there everything is working.
nvidia is the ones not a making a driver for it.
but to be fair they did describe wayland as terrible.
>>
>>79677467
Fake and gayland
>>
Wayland is the only way to make multiple monitors with different refresh rates work the way they are supposed to.
>>
>>79677661
>Wayland is the only way to make multiple monitors with different refresh rates work the way they are supposed to.
I literally have 3 monitors with 3 refresh-rates right now on X11 and they just work.
What are you going on about?
>>
>>79677467
>chromium works
last i checked only through xwayland, which is a shitty meme that wont scale properly on my hidpi monitors

pass
>>
>>79677620
Nvidia is the one not playing nice
>>
>>79677682
But I only have one monitor.
>>
>>79677645
>>79677703

Is it possible to make Wayland work using the mechanisms that NVIDIA provides? Yes
Are freetards refusing to do that? Yes

Look at the state of X11 with NVIDIA. Some compositors work flawlessly (e.g. kwin-lowlatency) whereas with others (e.g. picom) v-sync doesn't work at all.
Is this the fault of NVIDIA?
>>
>>79677620
>le evil nvidia
They aren't evil, they're just incompetent. If AMD near bankruptcy back in 2015 can create a good driver so can NVIDIA.
>>
>>79677682
Not that guy but I get tearing out the ass with my setup. How do you manage not to? Or do you just live with it?
>>
Please sirs, why not write the Wayland in Java. It will do the needful and I can get job on it to pay for dowry.
>>
>>79677682
>they just work
Tell me, at which refresh rate is every monitor set, and are the ones with a higher refresh rate displaying animations and smooth scrolling at the appropriate framerate?
>>
>>79677716
>Is this the fault of NVIDIA?
Of fucking course it is you dumb piece of shit. The nvidia driver is a broken shitfest.
>>
>>79677716
>Is it possible to make Wayland work using the mechanisms that NVIDIA provides? Yes
>Are freetards refusing to do that? Yes
you don't write a window server for a driver
you write a driver for a window server
end of story
we will not introduce breakages to the window server and segregation because one company decided they wanted to use their own thing instead of what is standard.
>>
>>79677740
Too low level.
But i bet you can create a decent if heavy WM with it
>>
>>79677767
This. EGLStreams is a buggy, incomplete and abandoned API. NVIDIA was going to create a new one but they got bored: https://github.com/cubanismo/allocator (last commit: 2018)
>>
File: dabbing boomer.jpg (50 KB, 577x1024)
50 KB
50 KB JPG
>>79677467
I'm using xsnow(6) RIGHT NOW and there's nothing that waytards can do to stop me.
>>
>>79677780
And so another pointless github project is created if only to get a pajeet Java code experience to pay for his dowry and porcelain shit chamber.
>>
I'd like to see the list of things that don't work.
>>
>>79677767
>The nvidia driver is a broken shitfest.
And yet the kwin-lowlatency devs managed to make it work.

How did they manage that?
> by doing nothing at all.
> NVIDIA doesn't require anything in particular for this to work. the buffer swap function seems to wait for VBlank.
(https://github.com/tildearrow/kwin-lowlatency/blob/Plasma/5.20/README.md#old-background)

wow so hard
>>
>>79677737
I'm that guy and that's the only way to achieve it. Either tearing or TearFree, which produces frame skipping at times.
>>
>>79677774
>you don't write a window server for a driver
>you write a driver for a window server

Actually you need to do both to have any hope of a working system.
>>
damn, feels awesome using radeon
>>
>>79677780
> too low level
Kek ... agreed. I only write my video drivers in Perl.
>>
>>79677737
No tearing on my AMD 5700XT.
I don't even use compositor.
Just FreeSync.

>>79677764
60, 144 and 165.
Yes, the higher refresh rates are actually higher refresh rates. You can easily see it by moving a mouse cursor.
>>
>>79677873
>You can easily see it by moving a mouse cursor.
I asked you this:
>displaying animations and smooth scrolling at the appropriate framerate
not the mouse cursor. Well?
>>
>>79677825
Or this, or a wildly successful project that pushes the wrong thing to the top and ends up being bought by google or something.
>>
>>79677873
Try opening a youtube video or testufo.com and moving the browser window between monitors. Enjoy the shitshow lmao
>>
>>79677891
I’ll do the logo
>>
Why is wayland such trash? First time I try it in 3 years and input still lags. I can't even type this stupid post without a delay.
>>
>>79677842
Damn. I 'd love to use a distro on my desktop as well (only use one on my laptop right now) but can't deal with every monitor but my primary tearing. Maybe once the XWayland acceleration patches from Nvidia are adopted.
>>
>xorg has chrome and steam
>wayland has gedit and gnome-calc
tough choice
>>
>>79677467
Don't quite get the point of this whole rant. Works fine on my machine with that reference WM of them. Zero problems literally.

Also:
>linux
>non-free drivers
Or am I getting something wrong here and nouveau is borked too?
>>
>>79677888
Tell me how to check the scrolling framerate and I'll post an update from home.
Cursor is the immediate way for me to check it usually.

>>79677907
Will check later.
Had no issues with MPV in the past.
>>
>>79677956
i can open steam on wayland no problem
>>
>>79678009
If you use firefox, go into about:flags and change layout.frame_rate to something like 165, for instance (since you use a 165hz monitor and firefox ain't all that smart). Try it out on your higher refresh monitor. Unless vsync is disabled (or you're using wayland), it will be locked at 60fps.

There's other ways to test that out too. If you disable all monitors except the high refresh one, you'll see a difference just by moving windows around between having only that display enabled and all of them. The mouse cursor consistently moves at the adequate framerate per monitor, but that's about it. Maybe some fullscreen applications too, and if I'm not mistaken, there is an environment variable only for nvidia GPUs that can select the default monitor for that framerate, but I don't have an nvidia GPU.
>>
>>79677835
Even using kwin-lowlatency, people are still having stuttering issues with nvidia. These particular devs thinking they've """solved""" one problem out of a hundred by ignoring half of all possible scenarios doesn't make the driver any less of an abortion.
>>
File: Xfce_logo.svg.png (331 KB, 1200x1200)
331 KB
331 KB PNG
Sorry but if this adorable little rodent doesn't work with Wayland then I'm staying put. XFCE is literally the only DE that doesn't make me physically sick.
>>
>>79677691
chromium --enable-features=UseOzonePlatform --ozone-platform=wayland
>>
>>79677620
It works on nvidia cards. As long as you don't get your drivers from nvidia, that is.
>>
>>79678412
Does it run Java?
>>
>>79677620
They were invited to the conference that came up with GBM but didn't show up, who gives a fuck what they have to say?
>>
>>79678429
Yes, I play Minecraft on Nouveau but it's slow.
>>
>>79677534
https://github.com/ifreund/river
>>
>>79678143
based
xfce is the only nonshit de work using
until we get a wfce or something similar it can be ignored
>>
>>79677907
>testufo.com
>>
>>79678132
>will be locked at 60fps
Here's my 165Hz monitor
>>
>>79679434
Honestly, not sure it's working.
With the manually set value it shows "165"Fps even on the 60Hz monitor.

If you have a better way of testing, I'll do it.
>>
>>79677467
>thousands of apps pruned for no apparent reason
>12 years the app replacements doesn't even reach the count of 100
kek, imagine if Windows did something like that to their users...
>>
nani? 99% or so of gtk, qt, sdl2 etc applications work natively, most of everything else works with a compat layer (which still runs better than native x btw). what are you on about?
>>
>>79678611
>Wayland requires you to implement half the system every time
>Your WM has no input configuration
>>
>>79680103
>Wayland requires you to implement half the system every time
It doesn't since wlroots became a thing but for some reason the big DEs refuse to use it.
>>
>>79680109
>but for some reason the big DEs refuse to use it.
This WM uses WLROOTS and still has to reimplement half the system.
>>
>>79679469
>>79679434
You should be able to tell when scrolling on any website. You don't need a test like that even. Since you have high refresh rate monitors I would assume that you can notice the difference between 60fps and twice/thrice that.
>>
>>79677467
How do input methods work in Wayland?
>>
>>79677518
Why settle for a dead and downstream GNOME fork when you can just use the upstream and have things work?
>>
>>79682127
because unlike gnome its good :)
>>
>>79677620
kill yourself Novideo faggot.
>>
>>79677956
chromium-browser --enable-features=UseOzonePlatform --ozone-platform=wayland


waylets btfo.
>>
>>79682346
Yes can’t beat the best. One thing GNOME lacks is competition as it’s too far ahead of the curve.
>>
>can't disable compositor
It's shit, is it even possible to make a plugin to disable it?
>>
>>79683068
A compositor is the software that puts all the pixel buffers together and sends it to drawing on the screen.
You can't disable it on xorg, when you do you are just using xorg's built in software compositor (which is shit).
Given what a compositor does, its also the perfect place to implement visual effects and double buffers for tearing, thus why people associate them with that.
>>
>>79682127
>GNOME fork
Unity is a Compiz plugin and has no relation to GNOME other than using GTK.
>>
>>79682651
But gnome is bad tho
>>
>>79683176
>which is shit
Why?
>>
>>79683176
You have no idea what you are talking about.
X11 without compositing uses the DDX driver backend with 2D hardware acceleration.
This means X11 draws directly to the frontbuffer modulu some memory fencing/dirty rectangle rendering.
>>
>>79685091
>2D hardware acceleration
No modern GPU has 2D hardware acceleration. Glamor is used for all open-source drivers which is OpenGL under the hood just like Wayland.
>>
>>79685117
>No modern GPU has 2D hardware acceleration.
Bullshit

>Glamor is used for all open-source drivers
Also bullshit
https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel
>>
>>79685277
>Bullshit
Not sure about Intel honestly but neither NVIDIA or AMD have 2D hardware acceleration for anything beyond a hardware cursor. You draw a flat 3D plane instead.
>xf86-video-intel
Barely receives updates and SNA (2D acceleration) is not supported by anything past Skylake.
>>
The interesting part (drawing directly to the front buffer without any synchronization) can be done by OpenGL or Vulkan but I'm not sure if Wayland's buffer model would allow it or if you would need to add a protocol extension.
>>
>>79677467
digimend drivers dont work on wayland, what to do?
>>
>>79677467
wayland is trash
x and arcan are vastly superior
>>
>>79685434
You have literally never used Arcan or Mir or any of these hypothetical Wayland alternatives.
>>
>>79685358
>Not sure about Intel
In other words, you were talking out of your ass. You ignore whole classes of modern embedded hardware as well.

>NVIDIA or AMD have 2D hardware acceleration for anything beyond a hardware cursor. You draw a flat 3D plane instead.
Nobody cares about shitty gayming cards.

>Barely receives updates and SNA (2D acceleration) is not supported by anything past Skylake.
Keep backpedaling.
>>
>>79685458
>you were talking out of your ass
Shouldn't have said "all". Though looking through this https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/tree/master/src/sna it seems the driver is just sending vertex rendering/memory management commands to the GPU, like OpenGL does. It doesn't have any fixed function 2D hardware like older GPUs.
>Nobody cares about shitty gayming cards.
Ryzen laptops are the best you can buy right now.
>Keep backpedaling.
It has been deprecated by Intel and only supports hardware acceleration up to Gen 9 (so Skylake). Distros switched to xf86-video-modesetting several years ago.
>>
>>79685549
not him but you got btfo very hard.
>>
>>79685616
How so? The thing he's talking about (drawing directly to the front buffer) doesn't even require fixed-function 2D hardware. You can do it with Vulkan using VK_KHR_display.
>>
>>79677703
I know, but I still need CUDA. Apparently NVidia devs are adding support for dmabuf :https://www.phoronix.com/scan.php?page=news_item&px=NVIDIA-DMA-BUF-Wayland-KDE

That would be a big step in making all Wayland compositors working for NVidia users.
>>
>>79685458
>embedded hardware
ahahhahahah
>Nobody cares about shitty gayming cards.
AHAHAHHAHAHHAHAHHAH

Holy fuck, how poor are you?
>>
>>79685549
>Shouldn't have said "all". Though looking through this https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/tree/master/src/sna it seems the driver is just sending vertex rendering/memory management commands to the GPU, like OpenGL does. It doesn't have any fixed function 2D hardware like older GPUs.
1. You wanted an open source driver that doesn't use Glamor. I gave you one and now you're just moving the goalposts.
2. That's still wrong. For example, it uses a blitter.
https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/blob/master/src/sna/sna_blt.c#L3675
and a video overlay
https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/blob/master/src/sna/sna_video_overlay.c
that aren't exposed in OpenGL.

>Ryzen laptops are the best you can buy right now.
20 rupees have been deposited into your account.

>It has been deprecated by Intel and only supports hardware acceleration up to Gen 9 (so Skylake).
You're moving the goalposts again. Gen10+ still uses the blitter.
https://gitlab.freedesktop.org/xorg/driver/xf86-video-intel/-/blob/master/src/sna/sna_render.c#L137
And it's still physically present in Gen 10+ btw.
https://01.org/sites/default/files/documentation/intel-gfx-prm-osrc-icllp-vol10-copyengine_0.pdf#page=35
>>
>>79677518
I miss Unity, bros...
>>
Let me know when I can run it without everything being much slower and eventually crashing and burning like the fucking shitty hack created by losers who didn't want to actually man up and do the job of maintaining Xorg like they were fucking supposed to so they started lying that Xorg is "BAWW UNMAINTAINABLE" and created some shitty little pile of trash that people still can't take seriously after FUCKING TWELVE GODDAMN YEARS OF DEVELOPMENT.
>>
>>79680109
>for some reason the big DEs refuse to use it
Nvidia, also dealing with Drew.
>>
File: waymeme.png (56 KB, 713x554)
56 KB
56 KB PNG
>>
Where is GTK2 with wayland support?
oh wait
>>
>>79686471
>2021
>Niggerviewer
>>79686496
>2021
>GTK2
There’s your problem
>>
>>79686513
>u dont need that bro
Take that gnomecock out of your mouth I can barely understand you
>>
>>79686527
GNOME Remote Desktop actually works, don’t bitch.
>>
>>79685434
Arcan looks interesting, but it's too early in development to be a daily driver.

It's developing faster than Wayland ever did though, not that it's a big threshold.
>>
>>79685358
>NVIDIA or AMD have 2D hardware acceleration for anything beyond a hardware cursor. You draw a flat 3D plane instead.
Nvidia's workstation graphics cards and Tegras have overlay planes in addition to a hardware cursor. All recent Nvidia cards have one or more copy engines that could be used to implement 2D blits, I don't have access to their DDX source so I'm not sure if they use it there, but I do know CUDA memcpys use it instead of running a kernel on the SMs.
>>
>>79686290
>that people still can't take seriously after FUCKING TWELVE GODDAMN YEARS OF DEVELOPMENT.
It's already the default in many popular distros stop seething.
>>
>>79677534
>bspwm
worst wm
>>
Basic desktop stuff works for me on gnome + wayland + nvidia nowadays. My games don't though which sucks and keeps me on xorg for now. My next GPU will be AMD anyway so I'm not too stressed about it.
>>
>>79685644
Doing a full modeset and tearing down the whole windowing system to work around its deficits isn't a solution. Why use Wayland at all at that point.
>>
>>79677467
DO NOT USE WAYLAND
>>
>>79683176
>you are just using xorg's built in software compositor
Do you even know what you are talking about?
Also, xcompmgr is pretty nice.
>>
Where are my hotkey daemons?
>>
waiting for a desktop that isn't gn*me
>>
>>79677467
>https://arewewaylandyet.com/
>tiling compositors
>sway
tiling compositors guys
>>
>>79685277
>>79685458
>>79685864
imagine being so confrontational when a person is calmly presenting some points lol 4chan has fried your fucking brain, go seek a mental health professional ASAP
>>
>>79684318
Unity is literally an alternative shell for GNOME. It was never its own DE, this is just bullshit millennial revisionism spread by contrarians who didn't even like it at the time.
>>
>>79677467
The default desktop environment of the GNU project already uses Wayland by default, so yes, we are Wayland.

You can deny it as much as you want, but nobody cares about your autismwm bubble.
>>
>>79680109
Because wlroots fucking sucks.
It was made by a bunch of autists who only wanted to appropriate Wayland development, hell, just look at Devault's "Wayland book", which is just copied from the official Wayland documentation.

wlroots and Devault are literally the "I made this" meme reified. They genuinely believe they were ever relevant and the driving force behind Wayland adoption, but the truth is, nobody ever gave a shit about them.

As much as you want to deny it, The only reason anyone nowadays gives a shit about Wayland is GNOME. Period.
>>
>>79685677
>i need cuda
If you develop on the same machine you use as a desktop, you're doing it wrong.

Not like you're not LARPing anyway.
>>
>>79677956
>xorg has botnet, wayland has free software
>>
>>79677645
>they did describe wayland as terrible
Source?
>>
>>79677467
Wayland is shit that breaks literally everything and doesn't have the basic interoperability features I've come to expect from Unix systems running a GUI. I'm not touching it.

>>79677620
>>doesn't work on 80% of desktop GPUs
It works fine on Intel and AMD iGPUs. You aren't a delusional retard that thinks every Chromebook and office PC is running a 150W RGB gaymong GPU, right?
>>
>>79680135
wlroots is not an easy library to use. Flexibility and "ease of use" are diametrically opposed, and wlroots deliberately aims for the flexibility side.

>>79680109
>the big DEs refuse to use it
I never expected them to use it. The target was always sway, and then other smaller compositors.

>>79691688
>They genuinely believe they were ever relevant and the driving force behind Wayland adoption
When has anyone ever said that? I don't really give a shit about Wayland adoption; I just think it's better, so I use it, and it contributed to wlroots and other stuff so my own personal computing experience would be better.

Also, Drew stepped down as maintainer for wlroots and sway, because he wasn't really doing anything. emersion has been the driving force behing all of that shit for ages.
>>
>>79691939
>Flexibility and "ease of use" are diametrically opposed,
Lol.
>Xorg is terrible and unmaintainable, but allows you to do whatever and is super flexible.
So to fix this you have
>Wlroots is terrible and unusable, but is also way less flexible than Xorg.
Cool.
>>
File: image.webm (2.72 MB, 1920x1080)
2.72 MB
2.72 MB WEBM
>>79683176
X lets you draw directly to the frontbuffer without any off-screen pixel buffers. Just glFlush or another graphics API's equivalent. No compositing needed! You can see this most easily by observing the classic window trails when you ignore expose events.
The forced compositing in OS X, Windows 10, etc is broken by design. It wastes video memory and increases latency for software which doesn't need it.
It's obvious at this point that Wayland is another sabotage effort by IBM to kill off more perfectly working features from Linux and make it as unusable as possible.
>>
>>79691960
God, you're fucking retarded. Stop talking about shit you know nothing about.
I'm referring to library/API design, but it does apply more generally.
>>
File: 1608070559969.jpg (79 KB, 748x756)
79 KB
79 KB JPG
>>79686513
Wait you can't even run GTK2 programs on Wayland? Seriously? No wonder people aren't using it.
>>
>wayland's assmad "developers" who abandoned their responsibilities for maintaining Xorg with the excuse of "IT'S UNMAINTAINABLE!!" (it's not) are still posting on 4chan trying to gain support for their trash pet project that has gone nowhere in TWELVE FUCKING YEARS
Pathetic.
>>
>>79691990
>I'm referring to library/API design,
"It's just a protocol, bruh".

Point stays, wlroots is terrible not only for the end users, but for the developers, who were the supposed reason of ditching X11.
And it took just 12 years.
>>
>>79691977
You are aware that all OpenGL/Vulkan programs draw to off-screen buffers, right? The X server just copies that shit over later.

>No compositing needed!
The X server itself literally composites things. It's just that its baked in one is awful.

You all make this dumb distinction about compositors because X made a big deal about whether it happens out-of-process or in-process, but that has nothing to do with the core of what compositing actually is.
>>
>>79691889
>>>desktop GPUs
>Chromebook
>>
>>79677645
>the code is fine, it just needs this third party to bend over backwards for us
How about you Loonix niggers develop stable kernel APIs?
It's unacceptable to require specific drivers for specific programs, even if they were open source it would be a waste of time for everyone involved.
>>
>>79677467
So basically all that shit only works on an AMD or Intel GPU?
>>
>>79692119
>It's unacceptable to require specific drivers for specific programs
Says the Windows luser who needs updated graphics drivers every time a new game is released that's specifically for that game.
>>
>>79691699
90% of people here who "need CUDA" are mining memecoins and not doing any actual development.
>>
>>79692119
>How about you Loonix niggers develop stable kernel APIs?
The userspace APIs are extremely stable.
The internal ones are not, and there are very strong sentiments against making stable APIs, for various reasons, with maintainability being a big one.

The specific API that causes "issues" is GBM, because GPU allocation is inherently unportable, and there is no standard kernel API for it. GBM exists entirely in userspace as a wrapper over driver-specific allocations.

I could go into more detail about why GBM was not seen to be good enough at the time, Nvidia's solution, which is garbage for everybody else and was rejected, and how GBM/the kernel has improved since then.
>>
>>79692139
>loonix retard doesn't understand drivers, example two millions and seven.

>>79692202
>with maintainability being a big one.
Sure, it's not easy to make an actually good OS, but if you don't get the basics right the rest of the effort goes to waste.
>>
File: 1591408071650.png (727 KB, 1202x1202)
727 KB
727 KB PNG
Prediction: X will outlive Wayland.
>>
>>79692244
>>loonix retard doesn't understand drivers, example two millions and seven.
You obviously have no idea how drivers work on Linux either, or the open source development model either.
I can link some talks from Daniel Vetter (main DRM subsystem maintainer, i.e. graphics) but you won't watch them.
>>
>>79692308
not that guy, link them
>>
>>79692360
https://www.youtube.com/watch?v=Kkv0Su0OMq0
Probably the most accessible talk. His linux plumber conf talks are far more focused at kernel devs.
>>
>>79692017
>Using GTK2 programs at all
>>
>>79692258
Wayland is already dead
>>
>>79692430
most of my regular gui programs are gtk2
>>
>>79692308
Not the wintard but I'd watch them, link pls
>>
>>79677467
NVIDIA is shit anyway you shouldn't buy their cards.
>>
>>79692308
>You obviously have no idea how drivers work on Linux either
Nah faggot, "no u" is not an excuse for the provably false claim above.
Linux at the moment has program-specific GPU drivers for Nvidia as the driver works only with the X server, Windows does not as the drivers work with the OS and at most need updates for program-specific tweaks: simple as.
>>
>>79692589
so obvious that it hardly needs to be said
>>
>>79692639
I don't really care about any of this but you've got to be one hell of a retard to claim something being api specific means its program specific
>>
Okay which one of you schizos made this "tearing protocol" merge request?
https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/65
>>
>>79692054
Laptops are desktops with low power components and a built in battery and peripherals, and that's exactly how people use them.
>>
>>79692794
>but you've got to be one hell of a retard to claim something being api specific means its program specific
You've got to be one hell of a retard to claim otherwise, especially given the apis in question.
>>
File: image.webm (2.91 MB, 1920x1080)
2.91 MB
2.91 MB WEBM
>>79692047
Here's a single frame from the Wayland 'compositor killer' on GLX with single buffering. 15000000 iterations. Explain how the partially rendered frame is shown if it's just copied later.
I know IBM pays its employees well, but nobody's gonna for your marketing campaign.
>>
>>79692047
not him, but from what i understand, compositing in the computer sense means each window has it's own personal buffer, which are then composited onto the framebuffer
while no compositor means windows draw to the framebuffer directly potentially overwriting stuff 'underneath' them, moving something 'out of the way' then requires the window 'underneath' to draw part of it's contents again, non-composited windows only appear to be composed from individual images, it's fake, what you see is actually all there is, hidden window areas aren't there at all
>>
>>79677810
based
>>
>>79692017
They work through Xwayland
>>
>>79693634
>wayland
>no screen recording
>has to use a phone and record his screen like an absolute boomer
ye im not switching lol
>>
>>79694226
That's Xorg.
It's recorded with a phone because the GPU driver won't let the screen framebuffer be captured until the rendering is complete, so it wouldn't show the tiling. I don't have a capture card on this system.
>>
>>79694226
ah yes, the hordes of screen recorder professionals coming out of the woodwork
>>
>>79695014
>y-you don't really need it
Why?
>>
>>79695213
didn't know that was the implication of that post, either way screen recording works and is implemented securely.
>>
>>79692119
death to proprietary software apologists
you don't get paid to defend them, why bother?
>>
>>79693634
>Explain how the partially rendered frame is shown if it's just copied later.
That's because whatever program is drawing mandelbrot is coded incorrectly and is sending half-done frames.
>>
>>79692132
You mean that shit that works on every single vendor, besides Nvidia, that doesn't follow standards and pushes their own bullshit.
>>
>>79695934
There are no frames (plural). It's a single call to glFlush.
>>
I've looked at the site and the list is pathetic, it is nowhere ready anons. Just wait 10 more years and then it will be ready (actually no, it will still be half-ready while X will be dead and usual stuff sabotaged and gnomers being dickheads)
>>
>>79691939
I love it. You felt so threatened when people called out wlroots on their Wayland appropriation that you materialized again after months.

You aren't fooling anyone anymore.
>>
File: 1609029862227.webm (1.23 MB, 1342x1080)
1.23 MB
1.23 MB WEBM
This is what X11 shills think is acceptable.
>>
>>79696887
Then blame the driver or something, because the output on screen is absolutely not what a single flush would look like.
>>
>>79685434
>Arcan
Now THIS looks really interesting, thanks for bringing it up. Seems like it'd be really fun to play with.
>>
>>79697357
What else should front buffer rendering look like according to you on a tiler, post the conformance test it violates.
>>
>>79695882
>implemented securely
>in C
buahhahaahhaahahahahahahahah

no
>>
>>79693634
Is that the one I wrote? Seems to be the first result when I google it.
I can explain what it's supposed to be doing if you want me to. It's designed as a test case, not as a client that is showing off how "bad" Wayland is.

>on GLX with single buffering
It was written with EGL.

Sounds like a driver fuckup. Things break down when you're abusing the system as much as that program is.
>>
>>79697471
If you're flushing only once it should look like a black screen before the flush, because you've sent zero data to the framebuffer before that point.
>>
>>79698685
Yes, just with some trivial modifications to run under X with GLX.
Why would rendering to the front buffer be a driver fuck up? It's a feature. While this is obviously a joke shader, there are real world rendering workloads in the vfx and cad industries that are similarly slow and need to be shown on screen as soon as possible.
>>
>>79698890
There's wasn't any initial color clear, the contents are allowed to be undefined. It's just garbage left over from the terminal window before it resized.
>>
/g/'s fucking knowledge on what a display server/protocol could do is fucking insane. Threads like these think you guys know more about wayland/x than anything else
>>
Schizo Wayland thread
>>
>>79699332
Imagine using GLX in 2021. It probably means you're unironically using Xlib instead of xcb too.
Use EGL. It would've required less rewriting in this case too.

I guess it's someone neat actually seeing the GPU finish the drawing tile by tile, and you can even see the vertexes being drawn (the diagonal line) but it still looks like hot garbage from an actual usability perspective.

The point of that program is to show an exaggerated example of the issues with implicit synchronization. If the compositor is clever, then such a slow client wouldn't slow down the whole thing, but no compositor does what is necessary for that yet.



Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

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.