Bloated, slow, ugly, old. Only used by autistic Luddites.
Hyprland is better.
 Nothing Ever Happens Shirt $21.68 |
 Tip Your Landlord Shirt $21.68 |
Nothing Ever Happens Shirt $21.68 |
Xorg
Bloated, slow, ugly, old. Only used by autistic Luddites.
Hyprland is better.
Nothing Ever Happens Shirt $21.68 |
Tip Your Landlord Shirt $21.68 |
Nothing Ever Happens Shirt $21.68 |
Works on my machine
>tears
shrug, I never noticed it. don't even run a compositor. Don't see any reason to move to gayland.
>i have a genetic eye deformity and tearing doesn't exist for me
I'm sorry anon, have you considered lasik?
that's because nothing that matters tears, like video playback and games
If you only use one monitor then tearing won't be a problem since fullscreen applications (videos and games) don't tear and your desktop doesn't move enough to notice tearing. The timing even without a compositor is good enough to not have tearing most of the time.
what distro uses /usr/libexec/Xorg? redhat garbage?
I did my homework of configure it. No tears in my desktop, only beauty
Yeah most people don't want to do that and then find out that tearfree doesn't even work.
That's your machine. Not his machine
no
m-muh input lag, though!
>casually ignores wayland's forced vsync input lag
xorg and wayland are both bloat, use TUI only
TUI is bloat use punch cards
Use an abacus, you filthy peasant
How do you do this? Just act like every computer is a terminal?
obvious satire is obvious
If you use lynx or that shit that turns Firefox into fonts you could use a TUI for everything. No downside besides lower resolution pictures.
Wayland is a humiliation ritual.
>tech luddite uses stinkpad from 2008
>his applications lock up regularly
>they crash
have you tried not being poor?
Wayland crashes with highend DPI mice.
Wayland also has horrible latency and for over a decade they forced v-sync (LMFAOOOO)
>badly designed protocol behaves badly
>this is the user's fault
And then waytrannies wonder why nobody likes them.
Wasn't linux meant for poor people? If you're not poor, you use Windows or Mac anyway
dont worry using linux long enough will make you poor.
increasing the buffer size isn't even a solution
i can suspend an X program indefinitely without issue
Xorg solves the issue by using both resizable buffers and suspending input events to non-responsive programs.
This is also theoretically possible on Wayland, but every compositor must independently work around the issue on its own. And even Weston and mutter only half-solve the issue (they stop creating input events for non-responsive applications, HOWEVER the message buffer is still tiny and fixed-sized, so it can fill up anyway.)
The screen-tearing on X11 issue is about as common as the mouse lockup on Wayland issue.
It only occurs with very specific setups and most people will never notice. Most of the tech from Wayland that eliminated it has since been backported to X11 compositors who now use the same rendering model.
These arguments are asinine and non-technical, though honestly Wayland has no real technical advantages beyond the security theatre and admittedly dumping the ancient rendering a.p.i. that no one uses any more is all fine.
I recently found out there's an application that does forwarding on Wayland too and when I read about the architecture I was like “OF COURSE”. It's actually a very trivial solution but I don't know how performant it is though compared to how clean it is on X11 it's an absolute hack though and it has to run a server on both ends whereas with X11 it directly connects to the server on the other end forwarded over SSH.
Also, the “push to talk” criticisms are asinine which can easily be solved with a global hotkey and i.p.c. though I suppose it's annoying that the hotkey is always bound, whether the swoftare is active or not.
Real criticisms on Wayland that remain:
>No standardized way to as much as move a window around, non-standaridzed things specific to compositors exist for some compositors and not for others
>Good luck ever running say the Sway bar on GNOME or vice versa. Though admittedly GNOME's bar is not a separate process but most X11 environments do make all their components separate processes exactly so that they can be ran in other environments. good luck even making a standalone bar that can run anywhere on Wayland.
>Same with things like a notification daemon
>Wine is still shit and doesn't work properly and only the absolute minimum of functionality can be ported
>Screenshot a.p.i.'s are shit.
Hilariously, Freedesktop Mission statement in 2016 contained this:
>Users should have a maximum amount of choice in selecting the applications they wish to run. Users should not be limited to a certain subset of applications; ideally, even the components of the desktop environment (window manager, panel, file manager, etc.) would be interchangeable.
They since changed it to simply:
>Users should have a maximum amount of choice in selecting the applications they wish to run
Because it as constantly held in their face as a criticism on Wayland that it went against their own mission statement.
Vagueness is a weapon in politics, and Freedesktop is first and foremost a political organization, not a technical one. Lack of vagueness is a weapon in technical matters.
>as common as the mouse lockup on Wayland issue
lol?
Literally first time I've heard about this.
I hear about it a lot.
Essentially, the Wayland event queue is too small and very high DPS-mouses send too many events so it locks up at times.
It's kind of a typical Freedesktop thing: none of the developers there use high end high DPS gaming mouses obviously so they thought an event queue of that size would be fine and it turned out to not be for high enough DPS mice.
It doesn't lock up, it just crashes random apps if there happens to be any kind of CPU load or disk I/O (Wayland devs forgot that literally any memory access can potentially result in blocking disk I/O.)
Although Wayland does have another mouse-related issue where the mouse will become choppy. Idk whether that's related to high-polling mice or not, thoughbeit.
>there happens to be any kind of CPU load or disk I/O
I've run fairly intensive computation on both GPU and CPU, I have an aerox 5, I've never had an application crash. The one thing I do know about is that debuggers or execution pausing in general will definitely crash an application. That needs to be fixed.
>Bloated, slow, ugly, old.
Agreed
>Hyprland is better.
No, and that's the issue.
That somehow something new can't compete with 30 year old junk because it tries to turn your computer into a phone OS.
>No, and that's the issue.
Hyprland is better. It can do everything I could do on X but better. It's faster, it isn't a slow piece of shit like picom, and it doesn't tear.
So, can you ssh into a different machine and run a single window on that machine while displaying it on the machine you SSH from with it?
I said I could do. I couldn't do that anyways because I never saw the need to learn how.
Anon, you do
>ssh -C -X <user@host>
><whatever command to start graphical thing>
The program will now run on the processor of the remote computer and access it's filesystem and everything but it will be rendered on the local one.
And yes, it's actually secure, because of the use of -X instead of -Y which uses the XSECURITY extension, so it lives in it's own namespace and can't actually inspect the X server on the local machine. Using -Y runs it in ”trusted mode”, disabling this and giving it full access to the X server of the local machine.
As you might imagine, this is an incredibly useful tool. The remote machine doesn't even need to have an X-server running or installed for this to work, so basically the usual:
>Wayland is gimped as frick, but I only use the most basic, elementary functions of my display server so I don't notice.
This is why people don't want to switch to Wayland.
yeah x11 forwarding was like the main wtf for me when messing with wayland how could they not include this functionality
You can see why in this thread
All the Waylandgays are really like
>Wooow, I never use this useful functionality and I don't believe anyone ever uses it.
Explains a lot about how gimped Wayland is. This is how the Wayland devs work.
>As you might imagine, this is an incredibly useful tool.
Really? Have you used this practically? I work with RHEL (send help) and I've never heard of someone doing this before.
Yes I do. I have two machines in my house.
Imagine that I don't want to get up and go to the other to start a graphical application and see the results at times.
That people working at Red Hat never heard of this explains a lot about that company and how insane it is.
Yes I am. You people are crazy.
Tell me how it's done on Hyprland because I can't find anything on it.
>That people working at Red Hat never heard of this explains a lot about that company and how insane it is.
i think he just means that he works with RHEL, not that he works on it. there's no fricking way that someone who works at RH wouldn't know about x forwarding, right?
Right...
right...?
RIGHT?
>I HAVE NO IDEA WHAT XFCE IS OR DOES, SORRY.
>That people working at Red Hat never heard of this
I didn't say I work _at_ Red Hat, I work _with_ RHEL. If we really needed a graphical interface for something it would be provided as a web interface.
>If we really needed a graphical interface for something it would be provided as a web interface.
this guy
shows real use for X forwarding.
Of course, it was developed with main frames in mind. Decades ago you'd have wanted to run X applications remotely because you didn't have resources locally. Back then, X forwarding was faster too because applications actually used Xr which meant that the drawing instructions were sent to the client, whereas now things get rendered locally and the client gets a giant bitmap over the network, which is also a nightmare for graphics acceleration. The new model is definitely using HTTP for this kind of stuff. X forwarding fits into a middle ground that is both secure out of the box and simple to use. I'm not going to run X applications on a server, I'm not even going to install X on a server to begin with. So the original use case is pretty much dead. X forwarding happens to work great for desktop users and hobbyists though. None of the security nightmare of vnc or rdp, and none of the upfront cost of using a web interface.
>Of course, it was developed with main frames in mind
I mean yeah I got that right away.
>I'm not even going to install X on a server to begin with.
right
> forwarding happens to work great for desktop users and hobbyists though.
I mean, I'd still love a practical example, like an application or something that you've used it with. It seems like a rather overkill way to check the status of something.
NTA but I used X11 forwarding to use Wireshark on my router PC before
What I use particularly often is the system monitor
>But but just use top which is completely inferior to the graphical process monitor that you use that has far more convenient functionality
No.
Also, gparted is an common target for me for whatever reason, not so much to edit partitions but to view them.
>But but, just use a terminal interface partition viewer that doesn't display things nearly as clearly instead
No.
I kneel. Come to think of it how it implements it of course it exists and it's actually quite simple. So it just spawns a local window and forwards between it? It must run a fake copositor at the other side and can't be too efficient right?
At least with X11, it actually connects directly to the local X server, there is no spoofed window in between.
>I kneel. Come to think of it how it implements it of course it exists and it's actually quite simple. So it just spawns a local window and forwards between it? It must run a fake copositor at the other side and can't be too efficient right?
>At least with X11, it actually connects directly to the local X server, there is no spoofed window in between.
if wayland devs really want to win they'll find a way to get better remote performance. imagine if you could actually run a remote GUI application with client side graphics acceleration. I'll bet $1000 they'll never even try though.
>if wayland devs really want to win they'll find a way to get better remote performance. imagine if you could actually run a remote GUI application with client side graphics acceleration. I'll bet $1000 they'll never even try though.
it would be THE killer feature but you're right, they wouldn't do it.
But for checking the status of something like a task manager or most gui applications? It will work fine.
>I'll bet $1000 they'll never even try though.
I think there's a chance Valve takes a look at something like this in the future. It depends on how their hardware ecosystem evolves though.
>it just spawns a local window and forwards between it? It must run a fake copositor at the other side and can't be too efficient right?
Most likely. There'd definitely be latency introduced on any kind of large redraw, I doubt you could use it for something like a video game.
I don't think X11 forwarding works for that either but with the compression option it's surprisingly snappy.
Without the compression option it's extremely sluggish though.
Also, I just discovered a hilarious thing about it I didn't know. Right now I have xrandr set to adjust color temperature because it's late here so I minimized blue light a bit but the forwarded window completely ignores that for whatever reason, no idea why, when it's part of this X server.
>no idea why, when it's part of this X server
If X actually handles color temperature per-window that's wild. It's definitely something that surprises me too.
no i'm pretty sure that is on the layer between the server output and display. the color temperature is applied to everything sent to the monitor. or the monitor itself is changing the display but I don't think xrandr uses ddci for color temp
i mentioned in an earlier post cryptocurrency. I run a full node for a few different blockchains, I can use ssh to get a GUI without trusting any 3rd party applications and while still having the benefits of full transaction history. It's secure enough I am willing to use it for crypto. I would never open the control ports for an application to control the node daemon remotely, I'd never trust a web based front-end to the node daemons, and I'd never have my private keys on more than one machine. Also, of course it requires a lengthy password even after you open the GUI. That's an example of something I do regularly. I don't have to put my private keys on any other machine and I can access those wallets via ssh over a vpn.
nta but i use it semi-regularly where i work.
this thing exist on wayland
and no. You arent actually using it
https://gitlab.freedesktop.org/mstoeckl/waypipe
does that work on macos and windows? because I can use remote X11 programs on a mac or windows pc at work without resorting to vnc or rdp
>at work
Is this through the LAN at your office or something? Or are you forwarding it from your home? If it's the latter you have some really lax policies. Also how am I supposed to get this on my phone?
over a fiber connection with a vpn it works fine. no forwarding of anything.
It's been years since I did anything similar on a phone, but you can install a linux userland system on android and get X, and there are plenty of ssh clients that support it. Wireguard is great on android so if I ever did that again i'd add my android to a vpn with my desktop and connect that way.
You don't even need a vpn though. For personal use, opening port 22 and only allowing public key authentication is fine.
> muh security
> muh lets copy vnc and rdp those protocols are so much better than x forwarding
waylands are useful idiots
not him but I use it all the time. IMO a great modern use case is simply and securely using a graphical crypto wallet from a remote machine. No port forwarding, no vpn or anything required. You can also combine x forwarding with xvfb to keep a gui application persistent across multiple ssh logins. The wayland solutions are far less granular and far less secure as of today. Also, moronic, you either get the same security with less performance by using ssh and xwayland, or you drop security to get performance
hyprland is bloat and maintained by a troony
link their github
last things I heard about trans and hyprland is that they bully them and drew devault got steamed about it
https://blog.vaxry.net/articles/2023-hyprlandsCommunity
https://blog.vaxry.net/articles/2023-hyprlandCommunityChanges
https://github.com/hyprwm/Hyprland/blob/main/CODE_OF_CONDUCT.md
The funny thing is even though hyprland happily bent the knee, devault and his flunkies are still assblasted and refuse to interact with hyprland users.
>Don't personally target people
>Don't go off topic in the repo
>Discord has a no harassment rule too but you can still express any beliefs
It doesn't sound that bad to me honestly
I feel that vaxry is slightly up for joking about trans people but doesn't care about it enough to threaten his software project in any way, not going to pretend to know the guy though, I don't use the discord and I think software 'communities' are a waste of time
Is there not actually a trans maintainer then? Perhaps it was mentioned in those links but I did only skim them as I read them at the time this all happened too
To me, Devault is at literally who tier, I don't know what he does of value and if he doesn't want to interact with me I'm ok with that
seems like a libertarian lite to me after reading a bit further
https://blog.vaxry.net/articles/2023-inclusiveActivists
Pretty good take, that's usually how I look at software. I'm autistic and don't really care about politics.
as far as software is concerned, I can appreciate that position
my main thoughts about him are: he makes something useful (a not-moronic wayland implementation) and drew devault doesn't
nah, that's a meme
x11 is exactly what you would expect
>x11 is exactly what you would expect
I don't expect tearing and a compositor with a slow, broken glx backend. I expect one application to be able to tear (like a game) and for the rest of my screens to not.
>gui
>mouse
>needs to click buttons to feel he's using a computer
>can't live without seeing his favorite troony porn
>x-gay or gaylend
>his gf told him she doesn't feel well tonight so no sexo for the 2000x time.
tty is indeed not for you.
>implying some autist living in the tty will ever have sex
lol
lmao even
But enough about Windows
nvm it says fedora right on the shot. lmao the arch experience without even using arch! also it's probably pebkac since its literally a screenshot of tmux. didn't know how to save the screen to a file or get a screenshot via ssh i guess lol
alternatives? there are no
They ripped off the twitter logo. Frick xorg. Fricking unoriginal fricks.
here's a fun wayland fact nobody seems to ever bring up--applications under wayland are unable to remember their previous window position, so they will always open in the same spot. by contrast, windows running under x11 (and, by extension, xwayland) are able to restore their previous position. so you end up with moronic dogshit inconsistent behavior because this is yet another basic longstanding X11 feature that wayland trannies are unable (or unwilling) to implement
>fun wayland fact nobody seems to ever bring up--applications under wayland are unable to remember their previous window position
I am not very experienced in how wms work, but should not it be a part of a wm ?
>I am not very experienced in how wms work, but should not it be a part of a wm ?
there's an argument to be made for that, sure, but it comes back around to one of the central problems of wayland--every wayland compositor is going to have to duplicate effort and come up with their own hacky workarounds. i'm also not aware of a single wayland compositor that actually implements it. like with many things, the way X does it might not be perfect, but at least it provides a standardized way of doing it and it works.
>with their own hacky way
Rmembering the position of a window and saving it somewhere is not going to be "hacky" in any compositor. A literal child could handle that.
i don't want my compositor to keep track of all my browser window names hoping to see that exact window name again so it could put it in its old location
>Rmembering the position of a window and saving it somewhere is not going to be "hacky" in any compositor. A literal child could handle that.
with how moronic and hacky everything in wayland seems to be, i wouldn't be so sure. straightforward features always seem to require some kind of painful workaround, just ask the mpv devs
the mpv devs have always been insane
Also wasnt that primarily about gnome? The way they're going if wayland didn't exist it wouldn't matter because they'd be using a completely non-sstandard x
That generally annoyed me on X and I wrote rules to make sure floating windows opened in the center because every stupid ass app decided to do it and also not have a toggle for it.
>That generally annoyed me on X and I wrote rules to make sure floating windows opened in the center because every stupid ass app decided to do it and also not have a toggle for it.
i prefer the exact opposite behavior, i find it very annoying when every application opens centered (the default behavior on the plasma wayland session) since i generally have my windows arranged in a particular way and i want applications to open in the same spot they were in before. not sure about other desktops, but you can toggle that feature off in plasma's X11 session if you want.
>I generally have windows arranged in a specific way
I enforce stuff like that with my wm because if I move the window I don't want it to open up in that position the next time, i want a default layout that I usually use.
i'm just too lazy to set rules for every application, i'd prefer to just move a window where i want it and forget about it.
https://dudemanguy.github.io/blog/posts/2022-06-10-wayland-xorg/wayland-xorg.html
>link
Yeah I've read that. Hyprland is better than X. I don't really care about the musings of some schizo.
Why are you so insistent on shilling hyprland? Go back to your troon discord and dial 8.
This is a friendly conversation Anon, there's no need to get your jimmies in a rustle
>I'm just too lazy
Same, thats why I use a dynamic tiler, it just places everything in a good layout by default in the majority of circumstances. No having to worry about moving the window somewhere or having to moce it back or setting a million rules. It just works.
i used dwm for a while. after plasma 6 fricked everything up im considering moving back to it (or another tiler), but honestly i like having a mix of a keyboard and mouse driven environment, whereas most tiling window managers err primarily towards keyboard control and there's little in the way of an interface you can manipulate with a mouse (awesome is about the only exception i'm aware of, since it's basically a full fledged floating window manager along with being a tiler).
>I don't really care about the musings of some schizo.
i mean, there's plenty more examples than that of wayland's limitations being a pain to work around, which was the point i was trying to illustrate.
>Yeah I've read that. Hyprland is better than X.
i'm not even an X diehard per se, and hyprland was pretty neat when i last tried it, but between hyprland's extremely rapid development and wayland still not being a particularly polished/feature complete experience, i don't see much point in switching to it yet. half of the appeal of running a standalone window manager (to me, anyway) is that most of them are relatively stable and unchanging, whereas hyprland sees a new release nearly every week.
>i like having a mix of a keyboard and mouse driven environment
sometimes I will use my mouse to move windows around or resize them (if they're floating especially), I have anything that isn't specifically related to window positioning/resizing mapped to the left side of my keyboard so I can still use a mouse comfortably.
>wayland still not being a particularly polished/feature complete experience
I genuinely haven't come across something that I did in X that I haven't been able to do in Hyprland. I did have to set a single rule for xwayland windows to work past some weird focus issue in Ardour, but outside of that it's been great.
>hyprland sees a new release nearly every week
That's fair, it's definitely rapidly changing, you'd have to use sway if you wanted something stable, dwl might work as well, but I'm not familiar enough with it.
wow their website is x.org
ain't that something
Bloated by 90s standards = lean and efficient today
nothing works on wayland just yet. support on wayland is ugly atm. xorg otoh supports pretty much anything ootb. for me it's xorg. wayland peasants btfod.
i use wayland and it works perfectly. kde plasma 6 uses wayland by default. x will be irrelevant in the near future.
Don't bother they will just endlessly b***h about non-features and nitpick everything no matter how much better it gets.
>i use wayland and it works perfectly
Lmao, no it doesn't, stop lying. Everything opens in the top left corner on KDE wayland.
oh boo hoo move the fricking window. i use hyprland anyway.
I just use a real window system (X11)
seethe about it
Bloated, slow, ugly, old, but at least it works.
>Hyprland is better.
Hyprland is a compositor/WM, not a display server, dumb moron.
Hyprland is better.
Crazy how a compositor manages to beat out entire display server.
it's almost as if that compositor was relying on another display server, huh
Oh yeah, and uh, what process is that display server running as?
something moronic, unstable and unfinished called wayland
said wayland also relies on X11 for most of its shit because nobody programs for it so you have to run 90% of your programs through the xwayland compatibility layer. yes, even on wayland and hyprland you're running most of your shit through X11
I don't see a wayland process in my htop
I also don't use any apps that run through xwayland, I have self-respect.
>I don't see a wayland process in my htop
I don't have an X11 or xorg process in mine but that doesn't mean it's not running
That's crazy because I totally see one running in mine.
>I'm a jobless neet
Got it.
I wasn't aware that you needed xwayland to administrate RHEL servers.
>He thinks managing rhel servers is a real job
You're just a filthy jeet who got filtered by learning how to code. Also
>Implying that businesses will suffer Xwayland's/Waylands garbage for any workload that requires GUI
Yeah I'm sure business are going to be fricking ecstatic that they now need to teach their employees how to move their mouse at just the right speed to stop Wayland from filling up an events buffer and crashing workloads and potentially losing data. Wayland is for no life neets who live in the CLI (you have to because GUI's under wayland do not function) and do nothing productive, like you.
What do you do for work, anon? Considering you use GUI software to such a significant extent I'd assume you're either developing it and need a debugger, or you're into something else.
>Yeah I'm sure business are going to be fricking ecstatic that they now need to teach their employees how to move their mouse at just the right speed
I use an aerox 5 with a high polling rate and I've never had an application crash. The only issues I've seen surrounding this is when programs are non-responsive or they've been paused. This is an issue that must be fixed, that's not something I disagree with.
>The complexity of making a Wayland compositor is overstated.
I would say vastly overstated.
>it's not needed I guess due to how simple it is to do
A lot of window managers still use xlib and are fully synchronous, which is very easy to reason about in terms of programming. It's not very complex. There are of course a number of xcb window managers as well, which are asynchronous. Those tend to be quite hard to develop, and you'll probably find a similar level of complexity within them as you would a wayland compositor that does not opt to provide compositing effects.
facts
>works
Seethe harder waytard.
I still have no idea what the deal between x11 and wayland is. I thought operating systems were the only thing that mattered.
We can't b***h about pulseaudio anymore because pipewire is actually good so now it's this
>We can't b***h about pulseaudio anymore because pipewire is actually good
I mean we still got people b***hing about systemd all the time when it's good too
>ssh -X
GOAT, makes gaylanders seethe
>why all the hyprhomosexualry?
cause israelitetubers are shilling it
Wayland is shilled mostly by clueless amdrone gaymen kids.
Just go to moronix in any wayland, nvidia, intel and amd article and you'd see the same accounts on plebbit and disqus.
It's actually hilarious because all their arguments are not techincal but boil down to nvidia bad.
I'm convinced that waytroon is an intentinal sabotage by redhat and ms to kill the linux desktop.
> Bloated
Runs on a 486
> slow
Uncomposited X11 beats Wayland on every metric including power consumption, latency and throughput
> ugly
Want to see something ugly?
Look at this: https://wayland.app/protocols/wayland
Or this: https://github.com/emersion/hello-wayland/blob/master/main.c
> old
Yes? 40 years of backwards compatibility. Any X11 program running on a VAX can still connect to a Xserver running on a modern system.
Sway uses less resources than i3 and I'd bet any x wm.
It also has noticeable input latency.
Can you prove that or is this a 30Hz monitor thing?
I really don't think I have hallucinations at my age anon. It's one of the reasons I can't switch to sway.
There is no way that any uncomposited i3 is going to take more resources than Sway.
With a composited i3, it depends on the compositor used and it's settings which makes it hard to compare.
The compositor is like 90% of the resources if not more of a common desktop. If I turn on blurring behind windows on picom I can immediately feel the snappiness tank a little. Composition takes up so much more than everything else the rest doesn't matter.
this is i3, uncomposited, with a solid background color, being run from and xinit script.
This is sway run from the tty.
Now I am using rxvt-unicode here. RAM footprint might be accounted to that, but I don't think my vram usage should be higher? That seems kinda weird, right? Granted, I am also using the vulkan wlr backend. Still, according to a lot of people that run an environment without any kind of composition, like you, sway should be using significantly more resources. I've never seen in that in reality though, there's never been a significant difference from running sway over i3 without picom. Now when I use picom over lets say, swayfx, we probably shouldn't get into that. The performance differences are rather embarrassing on Picom's side, especially with the glx backend. It would be nice if they worked on that a bit more.
To be precise, I'm using foot in sway and urxvt in i3
>Sway uses less resources than i3
No it doesn't
see
xorg and i3 uses more resources, not even just in RAM or cycles, but vram as well. That's without ANY composition too. Sway uses less and compositing is built in. Gee, it's almost like there's technical benefits to it.
X11 can do *a lot* more than Sway though.
If you want to replace X11 you have to include at least pipewire, waypipe, dbus and more.
If you want to have (bad) backwards compatibility you also have to include Xwayland.
>If you want to replace X11 you have to include at least pipewire, waypipe, dbus and more.
Okay, and is how these handle their appropriate operation technically better or worse than how X does? Lets take something like OBS. In X, if a window is not visible, it cannot be rendered through the window capture. Newer implementations are able to nicely work around that, and also seem to just in general be better at doing so. Even when urxvt was visible on X I would sometimes just get nothing. Some games I couldn't even capture at all. But with wayland/pipewire, it just works every time. Of course, I have the annoyance of having to click an extra button when I capture something, but at least it actually works. I wonder if Arcan would be better on this front, seeing as it's basically the newer paradigm nicely wrapped into a single implementation.
>If you want to have (bad) backwards compatibility you also have to include Xwayland.
So far everything I've run through xwayland just works after using a single rule to handle issues with focus. I haven't noticed any issues with performance either.
> In X, if a window is not visible, it cannot be rendered through the window capture.
This is of course a lie. XGetImage() can be called also on hidden or partially hidden surfaces. If the Window is minimized and does not render it won't obviously work.
So whatever OBS or your specific configuration is doing messes things up. Not X11.
> So far everything I've run through xwayland just works
Window managers, xdotool, xrandr and much more simply doesn't work. Xwayland is barely suitable to run legacy gnome apps. That's about it.
>If the Window is minimized and does not render it won't obviously work.
Yes. If I have a window on a different workspace or virtual desktop in any window manager, it does not render. This isn't the case with pipewire/wayland. It will always render.
>Window managers, xdotool, xrandr and much more simply doesn't work.
That's okay, I don't need to run a window manager, xdotool or xrandr through xwayland. If I want dynamic monitor configuration, I'll use kanshi.
On X11 know that they are minimized and can stop rendering if they so choose which saves energy and resources.
Unlike on Wayland where an application has know way of knowing if it is minimized or not.
If you wan't to capture a minmized window map it with xdotools.
>On X11 know that they are minimized
This isn't about minimization. This is about moving applications to a different workspace or virtual desktop. Looking at windowmap in the manual, this doesn't appear to accomplish what I desire.
considering their 'for the lols' UI from showed 'drag and drop to composite, filter, record and share' 10 years ago ..
Screen recording is one thing, but I'm a streamer, OBS is a very useful application. Still, if I could manage it through scripting lua, I wouldn't mind that either, I like lua (if only it had types, but I think there's a typescript transpiler for it). From what I can tell this video doesn't really provide a working example of the use case I'm concerned with. Although I'm sure Arcan would be able to handle it.
I use Fennel to transpile into Lua. Durden does more than what I want and Cat9 is gold. I asked the dev a while back about how sharing would work with their Xorg fork back in December. He just said to wait for an article on "Bringing X along fro the ride" on the blog.
>Fennel
Looks nice but I may not use it as it's a lisp dialect, the amount of my work with dialects amount to the configuration file I have for emacs.
Seen teal? https://github.com/teal-language
Now that's fricking awesome.
> but I'm a streamer
explains this entire reply chain tbh
What is that supposed to mean? I think I've been fairly understanding and level in my replies.
lol not at all. Like every streamer ever, your starting position is that since it works for you, naysayers are lying until proven otherwise.
Not everywhere on the internet is your chat room. I find streamers don’t realize how they sound when they dismiss the opinions of others as lies without a thought to the impression that gives.
This is a great example of what I’m talking about.
>naysayers are lying until proven otherwise.
I never said that, though. I have strictly talked about my own perspective, I haven't tried to make any assumptions for anyone else.
Reread my post. I clearly indicate that streamers often don’t realize the issue.
Are those two posts yours or not?
One of them is asking for proof of something, which is not saying that it is outright wrong or inaccurate. I just asked for something to prove the statement. That's all. The other is a bet. It's not an objective statement.
Again, are they both your comments?
Can you show how they would be representative of what you're talking about if they were?
I can, but I’m not going through the effort if you’re just doing to deny making the comments.
So I’ll take your twice dodging the question as implicit confirmation that they are yours.
So the first comment, you make a statement with no proof as fact.
In the second, you immediately question the validity of the claim.
Clearly when it’s your opinion, no proof is necessary on initial presentation of the claim.
However when it’s someone you’re replying to, you want proof before entertaining the idea.
Clear double standard. I can use your little resource experiment with that other anon as another demonstration if you want. You were clearly very confident in the validity of your experiment without even considering how it could be wrong.
Then you tried to misdirect with the vram comments and got educated for it here
Pretty clear case of “my shit don’t stink”, a classic symptom of being a streamer.
>if you’re just doing to deny making the comments.
Read this again:
>I just asked for something
Implying that I asked for it, so it was my post.
>you immediately question the validity of the claim.
Yes, if I see a claim I may not think is valid, or do not know enough about, I will ask for proof. That's pretty natural.
>So the first comment, you make a statement with no proof as fact.
That's not the same as objectively stating that sway is faster. It's saying, "hey, I think this is probably true". The other person actually immediately dismisses my proposal, in a very objective way. Saying that it's an impossibility outright.
>Then you tried to misdirect with the vram comments and got educated for it here
That one is actually not my post.
My response to that is here:
and here
Now you’re just coping. Hilarious that you think
>but it was a bet!
Is an excuse when we both know it was a figure of speech.
My point remains the same: you have a ridiculous amount of confidence in your inexperienced opinion, but don’t have the humility to even consider extending that confidence to the word of anyone else.
As I said, streamer brain.
Also btw, vram consumption is completely irrelevant in this case. It was explained to you but you ignored it.
>Is an excuse when we both know it was a figure of speech.
This is actually a totally natural way for people to claim that something is likely probably. And it's used like that as a figure of speech. I can't think of many people that would look at that statement and think
>yes, this is the person objectively saying that this is going to be the case 100% of the time.
>vram consumption is completely irrelevant
No, it's not, if I'm running stable diffusion, it absolutely matters how much vram is allocated. Likewise, if I have three monitors, and I'm running a game on the primary monitor, it also matters how much vram is allocated.
>xorg and i3 uses more resources
Not on my machine. i3 + xorg is about 23MB. sway is roughly 30.
Interesting that you didn't provide any evidence of this. At the very least show top or htop.
why is this so unbelievable to you? htop/top is unreadable so I'm not going to bother. Here is i3 + xorg. Apparently it's a little more now so 27MIB.
>vram consumption
This is completely irrelevant. It's not a video game.
>This is completely irrelevant. It's not a video game.
Having as much available as possible for games is relevant, though.
>Having as much available as possible for games is relevant, though.
That's not how it works. Your window manager or compositor is not silently doing 3D graphics calls in the background. The compositor will do it when compositing obviously but that isn't relevant when most of them have direct scan out by now.
>Your window manager or compositor is not silently doing 3D graphics calls in the background
So? Memory is memory, whether or not 3D calls are made.
Somewhat, but for me it's more about stable diffusion. Every bit matters when you're using rocm.
This is sway. Barely more but you're running xwayland anyway so obviously it uses more memory. At this level, it barely fricking matters though.
>htop/top is unreadable
Holy skill issue batman.
>vram consumption is irrelevant
Yeah no.
>it barely matters
It was a much bigger difference for me
>you're running xwayland anyway
Can't wait to drop that.
You're going to be running xwayland forever. Perhaps one day sway will make it run on demand when needed but until there there's an entire xorg server running in the background in addition to the wayland one. The entire resource argument is asinine until the. And again your window manager is not a video game. It doesn't use vram.
>It doesn't use vram.
If I start i3 and sway with no status bar and a black screen what do you think the difference in consumption is going to by, and why do you think there would be a difference at all?
If I then open up several windows of the same application, and X uses more vram than my compositor (as well as cycles), or vice versa, what would you equate that to?
Whatever renders the black screen uses vram for that split second and then it stops. Applications might use vram not your window manager. The only exception is when compositing (which is negligible generally), so it's really bizarre that you apparently care about micro-vram gains when an x11 window manager with no compositing would save your gpu the most.
>an x11 window manager with no compositing would save your gpu the most.
But it doesn't in my case.
The passive usage is just higher in i3. These are not transient values, they stayed like that the entire time I was setting up the screenshot.
You have more programs rendering in your xorg screenshot moron. It's not i3/xorg that's using the vram.
has nothing to do with vram
Okay, so you're saying if I have only those two programs rendering the difference should be less?
i3 resources wind up through OpenGL and therefore in the VRAM
Here I have i3.
>You're going to be running xwayland forever.
Unless I just don't use xwayland and don't compile support for it. There's pretty much only one application preventing me from doing that at the moment.
I'd also like to see the vram consumption. These are after all measurements I provided and the ones concerned with.
>Bloated, slow
I can run a full Gnome 2 desktop environment on a Pentium MMX with 64MB of memory
Can Wayland do the same?
That's impressive, I wonder if being able to do that would have any technical consequences for newer hardware. I feel like the inherent ability to support such ancient hardware must mean optimizations could be held back for newer hardware. At least, it seems to be that way from my understanding of WINE, which you have to fork if you really want good performance.
Gay wand. Need I say more?
>64 bit-only libraries and binaries
>wayland only
>pipewire
>stack wm
>no java 8
holu fugg send helb i can't worg on combuter :DDD
Frankly, the fact that you're going into a character assassination arch because I'm a streamer says a lot more about you than me. I could easily point to a huge amount of posts on this board that are far more inflammatory than mine, probably some of your own too.
For kicks, this is i3 with picom and the glx backend using damage.
bump
oh, you use wayland?
okay, now show me your xwayland running your videogames waychud
stupid dumb wayland scum
It just works.
>oh you use hyprland
ok now show me you running videogames in windowed mode or copypasting into wine
screenshot kde dolphin, your default file manager
show the world how broken it is
the only way dolphin is a hyprland default is one $fileManager variable set in hyprland.conf
I don't play games so I can't do your other thing
>copypasting into wine
What did he mean by this?
>screenshot kde dolphin, your default file manager
Sorry but I use ranger because anything without vi bindings is aids.
what does mutilating your genitals have to do with a desktop environment using shit software without features just to roleplay being cool?
>without features
Show me your wm that can tear a single full screen application and leave everything on your other monitors untouched. Show me your wm that lets OBS capture windows on different workspaces or desktops. Show me a single wm on X that has as good of a plugin system as hyprland. Show me your shell, no not polybar, I mean the shell you wrote yourself with a nice widget system. Go ahead, I'll wait.
>10 plugins
There's ~20ish now, some under development and don't yet appear on the awesome list. Then you have a bunch of official tools or tools designed with hyprland in mind like hyprpaper, hyprpicker, hyprlock, hypridle. Then there's some unofficial ones like hyprfreeze that you can use to suspend games and other applications that can't normally be suspended. Hyprshot for screenshots (which is just a wrapper for grim/slurp to make it easier), hyprkeys, hypract which is kinda like kde activities. It's got a very healthy ecosystem for such a young project that gets better week by week.
https://github.com/hyprland-community/awesome-hyprland
>14 plugins
wow worth mentioning when discussing genital mutilation and a featureless desktop software for sure
>can't count
There's 16 on that page.
>20 plugins
WOW genital mutilation in a desktop environment sure changed my mind
Not sure if the opinion of someone who can't count should be relevant to anyone.
>still waiting to hear what zinger is in store for the reason behind a desktop environment software needing to support or demote genital mutilation
for me its:
xorg: works
wayland: works
for me it's:
xorg: kinda works
wayland: works
arcan: doesn't work but I want it to work
>You will NOT make a window manager unless you reimplement the display server and compositor
Theoretically, what's stopping someone from making a Wayland server that lets another daemon decorate and manage its windows?
Nothing really. You can overwrite core parts of hyprland with its plugin system so it's already got similar parts of that. There's already an official titlebar plugin.
Weston actually allows that, and to be fair all the libraries that exist make it very easy to implement a compositor nowadays, from scratch.
Porting an existing system to Wayland that wasn't designed from the ground up around those libraries is the issue.
This is why almost no Window manager was ported, but rather many made entirely new ones as “drop in replacements” [read: 70% of things don't work like everything to do with EWMH] which is actually easier somehow; they weren't even forks which should tell you something.
The complexity of making a Wayland compositor is overstated. I mean, it is complex, which is why libraries exist for it. The interesting thing is that there are almost no real libraries for X11 window managers except of course xlib and xcb which are far more generic, because it's not needed I guess due to how simple it is to do.
In any case, you're right on the idea of the complexity being overstated I think.
>there are almost no real libraries for X11 window managers
xcb is good enough, wlroots is really the primary library used for Wayland if you're going to develop an independent window manager, and then there's Smithay if you want rust. I don't think there ever will be many libraries. It's a pretty extreme niche to create a library for. I don't think we'll ever wind up in a situation where the number of display server libraries outnumbers the population of India like we have with the web.
>desktop window managers need 20+ plugins before they can support troony genital mutilation
they must want hyprland to support genital mutilation because it's basically an analogy for what they do to people, cutting off parts and making them nonfunctional
is the troony in the room with us right now anon?