Kicad devs: do not use Wayland
Posted by FriedHoen2@reddit | linux | View on Reddit | 416 comments
https://www.kicad.org/blog/2025/06/KiCad-and-Wayland-Support/
"These problems exist because Wayland’s design omits basic functionality that desktop applications for X11, Windows and macOS have relied on for decades—things like being able to position windows or warp the mouse cursor. This functionality was omitted by design, not oversight.
The fragmentation doesn’t help either. GNOME interprets protocols one way, KDE another way, and smaller compositors yet another way. As application developers, we can’t depend on a consistent implementation of various Wayland protocols and experimental extensions. Linux is already a small section of the KiCad userbase. Further fragmentation by window manager creates an unsustainable support burden. Most frustrating is that we can’t fix these problems ourselves. The issues live in Wayland protocols, window managers, and compositors. These are not things that we, as application developers, can code around or patch.
We are not the only application facing these challenges and we hope that the Wayland ecosystem will mature and develop a more balanced, consistent approach that allows applications to function effectively. But we are not there yet.
Recommendations for Users For Professional Use
If you use KiCad professionally or require a reliable, full-featured experience, we strongly recommend:
Use X11-based desktop environments such as:
XFCE with X11
KDE Plasma with X11
MATE
Traditional desktop environments that maintain X11 support
Install X11-compatible display managers like LightDM or KDM instead of GDM if your distribution defaults to Wayland-only
Choose distributions that maintain X11 support - some distributions are moving to Wayland-only configurations that may not meet your needs
FactoryOfShit@reddit
I'm so glad that Wayland prohibits all this functionality that "people relied on for decades".
Yes, it breaks things, but knowing that you're finally in full control of all the different windows and apps cannot just move their own window/pop up over other apps without permission/glitch up and make the other windows uninteractable by stealing focus/track every keypress you make anywhere is a huge win.
Why does KiCad require this functionality? What's the use case for forcing the position of your windows? If you care about and want to enforce the relative position of your windows - perhaps it should be a single window.
alexforencich@reddit
Seems like things like docking tool palettes might need this. And yeah I guess they can rewrite half the application to do it some other way, but they don't have the resources for that.
LvS@reddit
Yes, all the applications that open multiple toplevel windows and want to attach them to each other in fancy ways have a problem with Wayland.
Previously they only had a problem with average users because those aren't used to apps vomiting tons of toplevel windows onto their monitor. Which is why Gimp redid its UI to not do that anymore and as a result is now a lot better.
alexforencich@reddit
Eh. Both setups have various advantages and disadvantages. Stuffing everything in a single window can be clunky with lots of dead space taken up by the one big window that has to be big though to fit everything else. Not to mention theming differences between the system theme and the internal windows. And wasn't that rewrite in the works for several years? I think the kicad folks are focusing their limited resources on the core of the application, not wasting time rewriting stuff that generally works just fine.
LvS@reddit
"generally works just fine" is a weird way of saying "is entirely broken on Wayland".
Just like saying "focusing their limited resources on the core of the application" about a project that is busying themselves regularly releasing FUD about Wayland.
alexforencich@reddit
"generally works fine" on the previous standard X11. Unfortunately some programs need a lot of reworking to work on Wayland due to the design of Wayland. Sounds like they're frustrated about the situation that has been forced on them by the transition to Wayland.
LvS@reddit
They've known this for over a decade and could have used that time to slowly transition to a Wayland-compatible design - the one that people prefer.
But they chose to be frustrated and go on a vendetta instead.
alexforencich@reddit
Do they really prefer it? Personally I have had more issues with Wayland than with X11, generally due to things that Wayland intentionally didn't implement. Screen recording not working, remote desktop not working, etc. Flip side, multitouch only seems to work on Wayland. One particular machine I have had to switch back and forth several times because certain things only work on one or the other.
And this kind of attitude from the user base is one of the reasons I am moving away from doing fully open source stuff. As a dev, you only get complaints, bug reports, attitude, and maybe a few cents in donations despite putting in untold hours of work, and after a while you start to question why you got involved in the first place.
bwfiq@reddit
Ive literally never had an issue with screen recording and I use nvidia? just use pipewire. The Unix philosophy is for every app to do one thing well anyway
FattyDrake@reddit
These were issues with Wayland early on I take it? Honest question. I've been using Plasma Wayland for a year now and screen recording has just worked (OBS and Discord), RDP has just worked.
Or is this more of an issue with LTS and Debian releases? I know they're always a couple years out of date, so being on a rolling distro would be a vastly different experience.
I wonder if that's where part of the schism comes from. LTS releases which don't have 2+ years of fixes being compared with rolling distros that have all the latest features working. As it stands with the new Debian release, any fixes and improvements over the next 3 years will be absent from a portion of users.
burning_iceman@reddit
Many people seem unable to differentiate between Wayland (the core protocol) and the Wayland ecosystem. Wayland devs refused to include features into the core protocol which don't belong there. Instead they are supposed to be specified in their own separate protocols. That hasn't changed, the other protocols simply got specified and implemented.
So when it is claimed "Wayland will never do xyz." this is probably true: the core protocol will never implement it. But it will almost certainly be possible on a Wayland system via the additional wayland-protocols.
alexforencich@reddit
It has definitely been getting better, but new stuff is always going to have kinks here and there. And these issues were with I think the N-1 LTS release of Ubuntu, so I think probably 22.04. And it wasn't RDP, but some other commercial remote desktop software that simply did not work at all with Wayland. And even RDP is kind of a pain in the rear as you need to plug in the screen size when connecting otherwise you only see part of the screen, and I have had issues with windows appearing to be transparent when they aren't and the mouse and/or keyboard input randomly not working.
LvS@reddit
The transition of Gimp to single-window mode was pretty unanimously welcomed. There were a few people who hated it but it was a pretty small minority.
The same was true for all other kinds of applications - browsers, text editors, terminals, file managers come to mind - that all switched to single windows with tabs about 10-15 years ago.
alexforencich@reddit
FYI it's common for PCB editing tools to use a two window design, one for the schematic and one for the PCB, with the ability to jump between the two (select a component on the PCB and jump to it on the schematic, and vice versa). With two windows, you can put each window on a different monitor, even if they're not the same resolution. So a true single window design isn't really optimal.
Also, with professional software like CAD tools, it's common to build the system (hardware and software) around the piece of software in question, so professional users it doesn't matter if it requires X11 instead of Wayland, they'll just set up exactly what's needed for the software in question to work properly.
lmarcantonio@reddit
The CAD/CAM community is not your usual office application one. User interface conventions *don't matter at all*, the job is usually complex enough that learning another application is trivial if it eases it.
Heck, I'd pay to have *two or more* cursors on two different windows to go from one layer to another (like gimp clone, more or less, but a key to change the live one). When you work with many layers transparency isn't good enough (dimming is useful but only works on the current layer).
Not standard? who cares, if it's useful.
lmarcantonio@reddit
Nope, not prefer. It's like the current gnome trend, they are forcing their decision. The multiple root gimp is the best way to use it when you have a huge amount of screen space.
thecavac@reddit
I'm lead developer for a Point-of-sale (cash register) system running on Linux. Opening windows at specific positions is a basic requirement.
Which is why we are not planning any wayland support in the foreseeable future.
ronaldtrip@reddit
So, what are the plans for the future? Maintain your own distribution for the POS? The magical incantation "Just use X11" won't work much longer on mainstream distributions.
gib_me_gold@reddit
X11 is not moving anywhere anytime soon from any serious distro.
ronaldtrip@reddit
Where the attribute "serious" is only bestowed by you on distributions that keep carrying X11 as a top tier display system?
The powers that be (a.k.a. the organisations and people doing the actual work) are moving things away from X11 at an accelerating pace, but here we are, celebrating the everlasting life of a display system on life support.
gib_me_gold@reddit
Find me one enterprise system that runs Wayland
Jegahan@reddit
Are you being serious? Redhat, the biggest entreprise linux system, dropped X11 in the latest release (10) XD
ronaldtrip@reddit
Yeah, the "it doesn't right now" defense. Have fun with the ostrich politics. Your X11 everywhere world will be janked away soon enough.
Don't think I want that to happen to you. Quite the contrary, but it is happening. The ones paying for development have annouced it, so it is going to happen. No amount of wailing by people, who don't pay and don't chip code in, will stop this.
It is adapt or be caught with your pants around your ankles. Maybe last minute scrambling is fun. I don't know.
thecavac@reddit
"people, who don't pay and don't chip code in"
I do open source projects as well as closed source that runs on Linux (hey, i gotta eat). No, i can't contribute to every library and subsystem i use, there are only 24 hours per day. I *do* push for more Linux on the Desktop through my line of work.
I just hope that, in general, the people who make the Wayland project keep in mind that someone also has to pay (and put hours in) for all the third party software to get adapted, and start working on APIs that are long term stable.
With all that init/upstart/systemd BS that happened in the last decade or so, i decided long ago not to use those systems for auto-starting and managing my software services. These days, my own service manager gets started by a crontab "@reboot" action and manages all the rest. And it doesn't ever need "root" to do its stuff.
If Wayland turns out to be the way forward and it's sufficiently stable (so i don't need to constantly fiddle because everything changed again), fine, i'll use it.
If it turns out that it's a constantly changing thing and a general pain in the butt, i'm not above rolling my own X11 packages if needed, if that turns out to be less painful.
ronaldtrip@reddit
Seems like a lot of extra work, but you do you. systemd has been the standard for a long time now and there isn't anything on the horizon that might replace it.
When it comes to Wayland... It has been announced 12 years ago that this would be the way forward on Linux. Despite Canonical temporarily muddying the water with their Mir display system. The design of Wayland makes it possible to make it expand to meet the needs of its time. If it is not ready yet, it means it hasn't been made ready yet by the people who need certain features.
Sure, it is always possible to Frankenstein something together and party like it is 2009. The question here is when it shifts into writing your own weird, legacy inspired, boutique OS? How many modern parts of Linux need to be bypassed to keep things as they are?
FryBoyter@reddit
As far as I know, SUSE Linux Enterprise Desktop uses Wayland as the default.
gmes78@reddit
It's already gone in RHEL.
thecavac@reddit
We will see. My own software isn't the only software that doesn't "make Wayland a priority".
I don't see a rush to spend lots of time to support something that is still in the "move slow and f up things" stage of development. Once the thing is sufficiently stable and feature-complete, i will reasses things.
burning_iceman@reddit
Why would the app need to control the position rather than the compositor?
thecavac@reddit
Because those positions are configurable in the software. The USER actually has no control of the window positions as per design.
Most of the stuff runs in Kiosk mode on a multi-screen setup. To make sure the windows go fullscreen on the correct screen, they need to be positioned on that screen in the first place.
For context: A window opens on the left side of the screen instead of the right, and they call support. Point of Sale, especially for gastronomy, is a high speed/high stress business, with multiple users often working on the same cash register. (They switch users in a second by pluggin in their keyfob into a magnetic lock). There can't be an option for a user to move a window, this would slow down other users, that would mean slowing down for a second to actually find the button positions.
burning_iceman@reddit
None of this explains why this cannot be configured in the compositor (by you).
thecavac@reddit
I already have. I use XFCE on X11.
burning_iceman@reddit
Great! So there is no issue with configuring it in the compositor and therefore no issue in this regard with Wayland.
Ripdog@reddit
Why? What use cases does a POS program have which wouldn't be solved by just using one full-screen window and drawing whatever you want inside?
gmes78@reddit
Then they'd have to change their code! /s
thecavac@reddit
POS systems often have multiple screens with different resolutions. To go fullscreen on the correct screen, you still need to position the window.
Changing the code is not the problem, that's easily done. But i'm not going to chase half-baked and quarter-implemented "solutions" for Wayland that change on a weekly basis when X11 works perfectly for our usecase. Maybe if Wayland has "grown up" in another 3 decades... i'll be in retirement and Wayland will be in use on enough systems that it will face the same scrutiny and number of security issues found as any other program ;-)
gmes78@reddit
On X11.
On Wayland, you can just ask to fullscreen a surface on a specific output, see here.
Antique_Tap_8851@reddit
This is exactly what this rant boils down to. It's an application developer trying to do unbelievably terrible things with hackish individual window management and usage that shouldn't be done in the first place and then cries when they're forced to rewrite their application in a way where you actually use common sense and implement things in much better ways.
It's the typical "boohoo, something is dying for good reason and we can't stand it because we don't listen to reason" knee-jerk response that should be ignored. Wayland definitely had its growing pains for a while but right now it's mature enough and good enough for people to use, and if an application fails to work properly in it, then that's up to the application to fix, not Wayland.
lmarcantonio@reddit
It's a usability issue. Even cursor warping if 'deprecated' by all the major UI standards but if it can make me work faster who cares. There's an option for activating a tool on the part under the cursor without clicking on it. Why? it's faster. It's like vi vs usual editors.
Don't want the 'strange' behaviour? just turn it off in the preferences.
Think about how blender works. OTOH they "solved" the problem implementing a whole window manager inside the application window.
Just keep the X11 capabilities in Wayland, when it's done we'll switch to it.
gmes78@reddit
Wayland already supports cursor warping.
Reasonable_Ticket_84@reddit
Cursor warping was only just merged, after over a year of kicad trying to push it. It was only merged after someone was nudged to vote for it.
It still remains that it'll be another 2-3 years before its available in stable distros.
YellowOnion@reddit
I'm running Blender in wayland mode, right now on a stable version of sway, and it's cursor wraps fine, I've also been playing with a locked cursor in video games, with a second monitor, where I'd be quickly leaving the window multiple times over at the speed I move my 1600dpi mouse in FPS games, and it's worked fine for 2 years. I can't help but wonder if KiCad is looking for a 1:1 spec without ever considering there's alternatives solutions.
Reasonable_Ticket_84@reddit
You aren't getting cursor warping in Wayland unless you are using XWayland.
Reviewing the source: https://github.com/blender/blender/blob/19c15e0218d732869658a33946d32f21ac42fb69/intern/ghost/intern/GHOST_SystemWayland.cc#L444
;)
teohhanhui@reddit
https://gitlab.freedesktop.org/wayland/wayland-protocols/-/merge_requests/264
It is coming.
olejorgenb@reddit
This could be solved by compositors allowing users to disable certain functionality (even on an per app basis). Yes, some apps would refuse to support this properly, crashing if the operation was disabled, etc. But given that these concerns are actually important to many people this would solve itself by making such applications unpopular. And the user can always choose not to use the application.
Personally I agree that window positioning is something applications should not mess with without extremely good reasons.
FriedHoen2@reddit (OP)
You want impose your view on kicad devs. A display manager should stay neutral.
FactoryOfShit@reddit
That's not what "display manager" means.
How am I imposing "my views" on kicad devs? If the app wants to be in full control - it goes fullscreen. If the app displays a WINDOW - literally the entire purpose of the window is to be user-controllable. That user-controllability is enforced by Wayland compositors, this is a feature, not an oversight.
If an app requires certain components to be arranged in a way the app wants - the app is free to take over and use a single window and manage the layout by itself. Using multiple windows is a choice that means "I am giving the windowing system (and by extension, the user) control over how these components are positioned".
FriedHoen2@reddit (OP)
MacOS and Windows dont impose these restrictions. Are they stupid? Maybe. Or maybe not.
FattyDrake@reddit
They impose restrictions and deprecate features all the time, especially Apple. With each new version of Xcode they effectively obliterate all references to previous versions. Either you update your app or you lose access to their ecosystem. There's no authority on Linux outside the major distros, which are only now getting around to saying, "X11 is no longer useful. Update or get left behind."
You just don't notice it on other platforms because devs are heavily incentivized to put the effort into updating their frameworks and apps for them because that's where most of the audience is.
FriedHoen2@reddit (OP)
MacOS still works fine with multiwindow apps.
FattyDrake@reddit
Yeah, but iOS doesn't, and Wayland is designed not just for desktops but multiple interfaces.
I'm also getting the impression you don't use KiCad at all.
I've been using it on Wayland for over a year and other than minor nuisances they point out (some of what they mention isn't even a problem in XWayland) it works fine.
Sounds like you're just railing against Wayland rather than actually concerned about KiCad's usability.
FriedHoen2@reddit (OP)
Agree, Wayland is not designed with desktop in mind, it is my point.
FriedHoen2@reddit (OP)
It's not me, it's kicad devs who cant afford to by-pass Wayland deficencies.
MarzipanEven7336@reddit
No, there just to fucking incompetent to pass a single fucking value to the windows constructor, which would make it run in a single window.
FattyDrake@reddit
Then how am I able to use KiCad just fine? You still haven't answered the question. What specifically about KiCad in Wayland prevents you, personally, from using it?
Otherwise, why bring this up at all? Again, since it was brought up the other day.
FriedHoen2@reddit (OP)
I dont use kicad, I linked a post by kicad devs who are are tired of bug reports by users who complain because they use Wayland.
FactoryOfShit@reddit
They don't impose those restrictions, and that leads to a terrible user experience the moment you encounter a bug in the app that causes it to pop up over other apps or steal focus. I'm glad I'm not using Windows, yes.
Or, you know, the completely useless privacy feature Windows has, where it tells you if an app is using your mic, which only works if the app is using the new API, which most apps do not.
But it's not stupid that Windows doesn't add these restrictions. Windows has a MASSIVE ecosystem of applications and Microsoft spends millions to ensure backwards compatibility as much as possible. The Linux Desktop is still maturing, and it's very much worth it to break compatibility now instead of whenever "the year of the Linux Desktop" actually comes.
FriedHoen2@reddit (OP)
The year of Linux Desktop will not come if Linux makes apps developers life a mess.
jasisonee@reddit
The status quo isn't neutral. It's just as opinionated as anything else. By implementing a feature that has existed before you are imposing a certain view on application developers as well.
FriedHoen2@reddit (OP)
X11 doesnt impose. If you want use 1 window is ok. You want multiple windows? Still ok. You want both? Still ok.
No_Hovercraft_2643@reddit
if you don't want other windows setting themselves over you? you lost. if you don't want other user programs to read your input (for example, your dudo password), you lost
al_with_the_hair@reddit
That's not what display manager means...
FriedHoen2@reddit (OP)
You are right, I mean a display server.
Qweedo420@reddit
I kinda understand tbh
I maintain a daemon that needs to know the currently active application for certain features, and I had to make a different implementation for X11, Sway, Hyprland, Niri, and I couldn't make an implementation at all for KWin and Gnome because they don't expose a proper API for that (KWin requires using JavaScript and Gnome only allows Gnome extensions to know the active window), I think we really need a unified API that gives the same features on every compositor
Technically there's the
wlr-foreign-toplevel-management-unstable-v1protocol but not all compositors implement it and I think it's really uncomfortable to usecrusoe@reddit
They took it all away because of "security" and now keep adding it back because it's necessary to be actually useful.
hitsujiTMO@reddit
IIRC, the thought process is that one application should not have access to another applications windows as that's a security risk.
And not being able to control your own application window is about portability. For instance, I use an app that requires a restart in order to change scaling. However, I use two different scaling factors on my monitors. 125% on my 1440p and 200% on my 4k monitors. On X11 if I want to move the app from 1440p to 4k, it needs to restart, where as, it doesn't on Wayland as it's all handled by Wayland rather than the app.
spreetin@reddit
My guess is that what will end up happening is that some library will emerge that abstracts away this by managing modular interfaces towards each compositor, exposing a single unified API for external applications to use.
serverhorror@reddit
You mean like we have one that provides a unified dialog for "Open File" ... oh wait ...
LowOwl4312@reddit
Great idea, let's get started!
I'll make the logo
spreetin@reddit
Honestly, this whole thread actually got me started looking into what would be needed to make something like this work. I might actually play around with it a bit and see if I can get something working.
ChronographWR@reddit
Just help xlibre
feldoneq2wire@reddit
Xkcd has a cartoon about this. Good luck!
YellowOnion@reddit
wayland-scanner generates C code from the wayland specs, including experimental extensions, it quite literally does exactly this. and because the spec is just xml, it also allows you to generate native bindings in your favorite language without having to deal with C middle ware.
asmx85@reddit
What should we name this library? I think something like "cross..." because it works across all these implementations. We should focus on the e.g. 12 most used ones. A cool nerdy abbreviation for "cross" is "x" so what about X12 ?
__ali1234__@reddit
How about Window Navigator Construction Kit, or libwnck for short?
gadjio99@reddit
No. X12 implies it's the next version of X11.
MeatSafeMurderer@reddit
wooosh
Reasonable_Ticket_84@reddit
This is a yo dawg joke right?
DividedContinuity@reddit
There is an xkcd fir this https://xkcd.com/927
dreamscached@reddit
They're not making yet another thing, they're making an abstraction level to encompass all of them with the same API. Not a relevant xkcd imo.
feldoneq2wire@reddit
It's still funny and relevant.
dudleydidwrong@reddit
That is what XKCD is talking about.
zocker_160@reddit
That "same API" is a new additional API though.
rEded_dEViL@reddit
Right. And the we’re back where it all started, replicating X11
zocker_160@reddit
there is a key-remapping application which allows application specific profiles and tries to implement it for as many Wayland compositors as possible.
https://github.com/cyber-sushi/makima?tab=readme-ov-file#application-specific-bindings
A library abstracting that away would be cool, but I think you are in for a very painful ride.
dagbrown@reddit
Maybe they could add the ability to run apps on remote hosts too!
cybekRT@reddit
Waypipe worka good and I've got lower latency than with X forwarding. But you have to have it installed on both sides, so this is a minus.
eliacortesi02@reddit
!RemindMe 1 year
zinozAreNazis@reddit
Ah bandages instead of fixing the root cause. Classic.
zquestz@reddit
Let me know when this exists.
olejorgenb@reddit
Gnome seems ridiculously hostile towards any generic protocol extensions. And their rationale does not seem to be primarily security, but rather that they don't want third party applications to implement what they see as the "Desktop domain"...
zocker_160@reddit
Sounds familiar.....same issue as with tray icons.
rockymega@reddit
Must... kill... GNOME...
Ullebe1@reddit
The 1.45 release of Wayland Protocols has a protocol for mouse warping in staging, so that shouldn't be a problem for very long.
I'm not sure what the status of something for window positioning is, but I imagine they can look at what the Wayland driver for Wine is doing, as IIRC that also had the issue that basically everything in Windows is a window and it needed to be able to position them correctly relative to each other.
tonymurray@reddit
Window position is going to be difficult. Wayland was explicitly designed to only allow the compositor to position windows.
Say for example, you are logging into something in a browser and another app draws over the browser and intercepts you click, and credentials. This is an example of a reason not to allow programs to position themselves.
There is also no global coordinate system in Wayland.
FattyDrake@reddit
It looks like the Wayland session restore protocol takes care of this. It saves the positions of the windows when an app is closed and restores them on launch. Meaning KiCad will not have to worry about this at all when it goes live.
tes_kitty@reddit
So, Wayland currently doesn't allow me to specify 'open a windows at (x,y)'? Who thought omitting that was a good idea?
So, an app will open where it was when I closed it? What if I had moved the window to the side, almost out of the screen and then closed it when I decided it was no longer needed. Will it open in the same position? Hopefully not!
tonymurray@reddit
FYI, there is no global coordinate system in Wayland. Sure compositors have them, but they can implement them as they see fit and may differ. There is no way for a Wayland client to know an x,y position of their window(s)
All compositors I have used are smart enough to avoid your scenario.
tes_kitty@reddit
Not sure if that's a good idea.
Does an equivalent of 'xmag' exist under Wayland?
tonymurray@reddit
They just have surfaces they draw on to. Not sure why they would care where the surfaces are..
Sure, Meta+Ctrl+scroll
tes_kitty@reddit
Because if an application has more than one window, it wants to be able to open the second (3rd, and so on...) near the first one? Very useful on multimonitor setups.
And that behaves the same as xmag? Telling me the X and Y coordinates plus the RGB values of the pixel I point to? I use xmag mostly to find out the RGB values of a given pixel.
tonymurray@reddit
No, it is a magnifier. There are color dropper tools on Wayland.
The application doesn't need the coordinates to request a window open to its right. But they are still working out details for positioning multiple top level window applications. I suggest you check the discussion for this on the Wayland repository as you seem to be interested.
tes_kitty@reddit
Wayland has been in the works for how long? Sounds a bit late to start working on such important details, that's something that should have been included from the start. Reminds me about IPv6 and DHCP.
tonymurray@reddit
If it is so important, how is Wayland adoption so high without it?
FattyDrake@reddit
Oddly enough, by allowing an app to specify x,y, you can have a window appear offscreen. Whereas the compositor is in control, it can reposition it to be more visible in case it ends up offscreen (like when turning a 2nd monitor off) or close to offscreen where controls can't be reached like you mentioned.
Currently in KDE, under System Settings->Window Management, I can specify coordinates for any application, so whether it defines it or not. In that respect, the Wayland way offers a lot more control for the user. Even now, without the session restore protocol, I can set up KiCad's windows to appear in specific places whenever I launch it if I want.
somethingrelevant@reddit
you can definitely allow apps to place windows while also preventing them from placing them offscreen though
you can definitely do this in xorg just fine
tes_kitty@reddit
How do you handle this if you run multiple instances of the same application. xterm or any other terminal emulator comes to mind. Do they all open in the same spot and then you have to move them or can you specify where each window will open?
FattyDrake@reddit
That's a good question! I only ever use it for individual apps that I want the position remembered or on multiple desktops when launched, etc.
I tried it with Konsole, and it indeed opens up in the same spot. You can base it off window title if you want so if each instance of the same app has a different window title like Untitled 1, Untitled 2, etc. you can adjust it. (Haven't played with the scripts much, maybe there's something there.) But Konsole has the same title for each new window. Possibly a bug! I'll gather more information and ask on the KDE forum. Definitely something to be aware of before they implement the protocol.
tes_kitty@reddit
So when you have your screen full of terminal windows, log out and back in again, do they appear in the same position as before or all in one spot on top of each other?
zhurai@reddit
What if I don't want to see the controls? (I don't know/doubt this is entirely applicable to KiCad, I don't use it)
KDE does not let me do this (e.g. in my inputless screen that I purposely have multiple miniture windows in a dashboard esque way to monitor/handle multiple things in a single monitor -- pretty useful for monitoring things)
KDE does help by removing the titlebar/frame, but some of the upper elements still don't get hidden in this way, and I'm not able to use negative x/y to place the screen slightly offscreen. (but keep the main important part of the application visible)
AyimaPetalFlower@reddit
super+click
zhurai@reddit
That does not let the window go to negative y as that just maximizes it, and is not a window rule that forces said window to stay there no matter what (set position/size + force)
Additionally, it does not let me to decide this based on a unique window title, e.g. to move a specific Firefox window partially off the screen (hiding the url bar, that's already put to a smaller ui for other windows)
For example, if I have a screen near the top that is looking at a grafana dashboard or icinga dashboard, and I do not plan to manually change the url (but using the same profile) so I'd like to move that off the monitor viewing space.
However, setting a negative x/y causes the window to completely disappear.
AyimaPetalFlower@reddit
super+click on my machine lets you move windows without using the titlebar and move content off screen unless you're talking about something else, I've done it before. right click also works for resizing on most desktops I've used including kde. Windows doesn't let me do this and it drives me insane since I'm so used to not having to click titlebars.
I'm honestly not entirely sure what you're referring to but you can also have your compositor lie to apps and tell it to have the fullscreen hint and use "fullscreen" firefox that auto hides the titlebar but functions as a normal window.
Is this what you're talking about or something else?
zhurai@reddit
Hi. Thanks for the response, but I think there's a bit of misunderstanding in what I am doing/trying to do in my setup.
I am using window rules to: Automatically move and force part of the content off screen, and then also force the window to be a certain size (no matter what the window thinks it's minimum should be) as this is a computer that is keyboard/mouseless
(Basically, this is a computer that does not have a set of keyboard/mouse, and is only controlled by deskflow/input-leap/synergy from my main computer, so you can think of it almost like a kiosk that is designed to monitor things without any user input, and unless I want to have 3 screens for it, I want to remove as much from the window that doesn't matter for my monitoring usage.)
Basically KDE does most of this, but doesn't let me use negative position values (as the window itself just disappears if you use a window rule with a negative position value), but at least removes the titlebars and everything else listed. (Unfortunately though there's still some things that still take up a lot of space)
Thank you for the Fullscreen hint idea, I'll see if it helps with this.
lmarcantonio@reddit
Kicad needs to close and reopen in the same place during execution, wouldn't work. Also there are *lots* of use case for having absolute positioned windows, like 'stuck' floating toolbars (gimp users know what I'm talking of)
TheOneTrueTrench@reddit
Yeah, I'm reading the list of "defects" in Wayland that conflict with KiCAD, and I'm just seeing a list of major security flaws in X11 that Wayland has fixed, and the KiCAD developers are just upset that the security flaws they chose to rely on are finally being fixed.
chrisoboe@reddit
Not only a security problem but also horrible ux.
I want my wm/compositor to place my windows in a unified way.
I don't want that each application does it's own windows placement where everything behaves completely different depending on the software I use.
lmarcantonio@reddit
Warping in kicad is a user preference and windows pre-placement can be a nuisance (I personally patched away the shim in my git).
Anyway it's not necessarily a safety issue if it's not interacting with other processes (yep, the *recommended* way to use kicad is as a single process, and I hate it).
I think that *not* supporting X11 idioms however is a step back. The safety issues could eventually be solved with some kind of capability flags (like... can this program warp the mouse? yes/no)
TheOneTrueTrench@reddit
If KiCAD can do it on X11, any program can do it, and it's REALLY easy to use stuff like that to manipulate user interaction to do whatever you want.
... that's called a Wayland protocol. You just suggested doing things the Wayland way. And it exists.
feldoneq2wire@reddit
Once you install a piece of software, the ship has sailed on that software betraying you. Disallowing an app from repositioning its own windows is how you get zero games or pro software on Linux.
TheOneTrueTrench@reddit
That's hardly true, or containerization, SELinux, AppArmor, and the entire security infrastructure of every major OS wouldn't exist
feldoneq2wire@reddit
"software that you rightfully download and install shouldn't be able to do anything"
Might as well install Windows 11.
RoboticInterface@reddit
What a straw man argument.
There's a reason why not everything runs as root.
People use Linux because they want to, but people also develop Linux because they want to. Turns out many developers care about security and working on a modern code stack.
You are free to run your system however you want to with the tools available, but don't expect developers to bend to your preferences.
feldoneq2wire@reddit
An app being able to reposition its own windows isn't asking for root access. If you think apps should have very restrictive access to the OS then I'm sorry but you support the Microsoft/Apple/Google model. That's not a straw man at all.
tonymurray@reddit
You clearly don't understand the security issue.
RoboticInterface@reddit
"software that you rightfully download and install shouldn't be able to do anything"
As you have missed it, this is the straw man I am calling out with my "there is a reason why not everything runs as root" comment. Creating a better architecture for security is not stopping programs from running, it's giving more control to the owner of the system by letting them define what programs are allowed to access.
"An app being able to reposition its own windows isn't asking for root access." My point is that security has always been something that is important and being improved on with Linux. Reasonably, why should an app be able to move its own window? That is the compositors job. And if there is a use case then an interface should be made and I should also have the ability to revoke that access.
"If you think apps should have very restrictive access to the OS..."
I DO think apps should have very restrictive access to the OS, or however much I tell them to have. I care about having control of my OS, and I want the power to be able to allow or deny what an application can or can't do. The difference from MS/Apple/Google is that I own every part of my system, and I am the one responsible for it. Every Linux user should be able to choose how much security they want, that is a good thing.
nightblackdragon@reddit
Not very good option for user. Remember how users disabled User Account Control in Windows Vista because it was too annoying?
lmarcantonio@reddit
Android does exactly the same thing. For some things the user has actually to confirm the permission on the control panel. The Linux kernel actually has capabilities to subset root privileges; the "allow only" ACL policy is actually the only one it understand (IIRC they refused to add the NTFS-like permissions to NFS4 just because they go against that)
Also: you can't remove features just because the user can't notice, for example, that it's browser window go inactive when she types the password. These days people paste bad stuff from internet into the run box, should they remove the run box? (yes, there are GPOs to do that, I know...)
FriedHoen2@reddit (OP)
They are not interested in workaround Wayland defects.
Ullebe1@reddit
I guess that's just super unfortunate for Kicads users then.
FriedHoen2@reddit (OP)
Kicad users will continue to use X11
TheFr0sk@reddit
Or just stop using Kicad like myself
mrdeworde@reddit
Sincere question here, what's KiCad's main alternative in the FOSS/Linux space?
FattyDrake@reddit
Horizon EDA gets lots of great reviews and even uses KiCad's routing model.
mrdeworde@reddit
Thanks for satisfying my curiosity
TheFr0sk@reddit
Not fully open source but I use the browser version of EasyEDA
wintrmt3@reddit
X11 is dead, KiCAD will die too if they don't change this idiotic stance.
FriedHoen2@reddit (OP)
KiCAD is a multiplatform app. It works on Windows and MacOS too. To it will not die.
pppjurac@reddit
This spells like further decline and abandonment of linux port.
jcelerier@reddit
as a developer of an app where not having mouse warping makes it *infuriating* to use, what's the ETA until the average Debian user can use my app properly on wayland in at least the mainstream compositors (mutter, kwin, wlroots ones)?
SufficientlyAnnoyed@reddit
I love Debian, however, maybe you need to use something that's a little faster on new software versions than Debian?
jbicha@reddit
2.5 years
burning_iceman@reddit
Much faster actually, since the average Debian user could switch distro rather than wait.
Wonderful_Turnip8556@reddit
agreed
gmes78@reddit
Debian 14.
Neomee@reddit
Like... There are so many complex applications working just fine on Wayland... Kicad is any kind more special or that is just unwillingness?
burning_iceman@reddit
They're unwilling to fix bugs on Wayland and then complain about a buggy experience on Wayland.
gib_me_gold@reddit
It is impossible to fix the bugs without re-doing half of their application, and the bugs stem from issues with Wayland, NOT issues with their own software.
burning_iceman@reddit
Wrong, mostly they stem from issues with their software or their toolkit. The only two issues with Wayland is pointer warping (newly available) and session restore (coming soon).
RoryYamm@reddit
Their software works on Mac.
It works on Windows.
It works on X11, which means Linux and BSD.
It doesn't work on Wayland, because the protocols don't support the things it does with its windows and UX. Specifically, the ability to place its OWN windows in a specific place, as dictated by the program itself.
If it works everywhere except Wayland, and XWayland can't even save it, then the problem is Wayland. The session restore protocol won't even fix it. What KiCAD needs is the ability to say 'PLACE THIS WINDOW AT THESE COORDINATES'. but apparently, we need to keep that out of the protocols because VR or some bullshit.
burning_iceman@reddit
Firstly, it works fine on XWayland.
Secondly, as others have pointed out, while they say they need 'PLACE THIS WINDOW AT THESE COORDINATES', they don't actually. They could solve it differently, but they've restricted themselves unnecessarily to one preferred solution. There's good reasons to not allow regular programs to decide the position. Besides the security issues, individual programs don't take the variations of screen setups into account properly. The compositor is better set up to place the windows correctly.
Regardless, almost all of the issues in the article are caused by their (or their toolkit's) poor support of Wayland.
_chococat_@reddit
As the article says and as a user of Kicad, I want to design circuits and circuit boards. Anything that slows that down is an annoyance. Also, as has been mentioned, a lot of the issues stem from wxWindows not being brought up to date with Wayland, hence the developer's suggestions to contribute upstream.
thunderbird32@reddit
Considering it won't be long before all the big commercial Linux distros are Wayland-only this is a bit laughable. RHEL 10 is dropping support, and I expect it won't be long before the others follow suit.
DoucheEnrique@reddit
Weird take. Using Linux professionally does not imply you have to use enterprise Linux. Everyone using Linux to earn money is using it professionally. For example freelance illustrators like David Revoy who apparently is using Debian.
dajigo@reddit
FreeBSD is the way.
Technical_Strike_356@reddit
And what do you expect that they do? The page makes it pretty clear that Wayland is not ready for KiCad-quality software.
FriedHoen2@reddit (OP)
One can use RHEL 9 until 2032. Maybe Wayland will be ready.
ranixon@reddit
Cursor wraping was released recently
Windows positioning, still on the works, so still not working
akp55@reddit
so kinda some basics that we would all expect, but they are just being released.....
ranixon@reddit
Yes, sadly, wayland is slow. This is the problem of democracy and consensus, everyone should agree on it. Windows and Mac would never have this problem becouse they can do whatever they want, specially Mac.
grem75@reddit
Canonical tried to go the dictator route with Mir too. They had something functional quickly and if they wanted to add something they just did it. That doesn't really work in the open source space when you're talking about something as fundamental as a display server, so no one else was really interested in doing much with Mir.
dagbrown@reddit
They also tried to go the dictator route with Upstart. That went well.
grem75@reddit
Went a bit better than Mir at least. There was some adoption by others, even RHEL 6 used it. I think ChromeOS still uses it. At least Upstart came before systemd, which gave it some chance.
Saxasaurus@reddit
which makes it arguably more successful (more users) than systemd in the desktop distro domain
Oerthling@reddit
Actually it did go fairly well. Upstart had wider support and several distros were adopting it before SystemD overtook all the alternatives.
I don't see what's wrong with Canonical taking the initiative in building an imorovee service manager.
Nor was it wrong for red Hat to them develop SystemD. Several solutions were competing at the time - SystemD won. A couple of alternatives are still around for the people who prefer those over SystemD for one reason or another.
I'd say the system works.
midorikuma42@reddit
For those, it seemed to work ok.
For display servers, the system doesn't seem to be working at all.
Oerthling@reddit
Why not?
X11 eventually ran into architectural problems . So the devs came up with a modernized concept and developed that for years in parallel. And there's even an adapter to run the old software on the new version.
It's a complicated big thing with many particulars, so it took quite a while, but we're entering the final stretch of the transition.
Wayland mostly works fine. Millions have been using it for years now.
So how is it "not working at all" - while it obviously works a lot.
burning_iceman@reddit
The issue was how they wanted to have an unreasonable amount of control via CLA and how they tried to use their influence in the technical advisory board to push it as the standard for Debian over the superior systemd. So political/legal issues rather than the initial idea.
Oerthling@reddit
They had just invested in a fine service manager: and with invested I don't just mean developer hours/money, I also mean the effort of the integration and the buy in from developers.
It's totally understandable that they weren't fond of immediately switching again. Plus, I agree that SystemD is ultimately the better solution, but we both know that SystemD had quite a bit of widespread controversy and plenty of people, outside Canonical, who very much disliked it (massive understatement detected).
And the way I remember things happening Ubuntu fairly quickly switched together with Debian.
midorikuma42@reddit
In retrospect, I wonder if the Linux community would have been better off just adopting Mir.
grem75@reddit
Mir had plenty of its own issues and was too focused on Canonical's personal goals.
CUPS at least has a simple goal everyone can agree on. It was also an existing project that was open source before Apple bought it.
ronaldtrip@reddit
Don't forget the asymmetrical licensing on Mir. Canonical having an irrevocal and perpetual license to relicense to whatever they want (through the CLA) and the rest of the world having to be content with the GPLv3. That is also a major reason that the rest of the world kept going with MIT licensed Wayland.
somethingrelevant@reddit
so far part of the problem with wayland seems to be that users generally agree certain things should be possible but the developers don't, so those things don't get implemented even though obvious use-cases exist. and of course the devs get to choose how the software they're developing works but it creates this big disconnect between what wayland has and what it actually needs
ranixon@reddit
The problem with the edge cases is, do you want to keep the protocol coherent or you want to add workarounds of workarounds?
Read the Windows positionig merge request and you will understand, specially the FAQ.
Awyls@reddit
I understand both sides of the argument, but something like this HAS to be slow. If you fuck up, its there forever and in the worst case scenario you end up with X11 where people can't even work on it.
Honestly, this is not as good as it looks. Just look at Mac's transition to Metal or ARM. Just a complete shit-show.
arthurno1@reddit
People can fork X11 if they want, and there exists other (commercial) servers that speak X11 protocol.
nightblackdragon@reddit
What's wrong with Mac transition to ARM?
mishrashutosh@reddit
Not sure about your last point. Mac's transition to ARM was as good as such a transition could possibly go. The first ARM Macs were released less than five years ago, and pretty much all popular Mac software is already ARM native.
mrlinkwii@reddit
their is slow and then theirs wayland slow which moves like an iceburg , pulse audio was replaced in 5 years not 20
Sataniel98@reddit
Microsoft has been re-releasing Vista with more or less zeitgeisty UI revisions and more high level standard software for the last 20 years. Same for MS Office. They don't have development speed issues because they don't develop anything big (other than the integration of their subscription services).
mrlinkwii@reddit
no , this is the problems with bikesheeding and telling usersd their use is bad and should be changed
HyperFurious@reddit
Wayland is slow but wayland developers are angried because some people think that X11 is not dead yet.
MyraidChickenSlayer@reddit
Because wayland is slow
Exponential_Rhythm@reddit
"After seventeen years in development, hopefully it will have been worth the wait."
SchighSchagh@reddit
Just imagine if X11 had received actual development over the past 17 years. All the security issues would've been fixed, VRR, HDR, vsync, all could've been implemented, and nothing would be broken.
Infamous_Process_620@reddit
idiotic take
"All the security issues would've been fixed" a lot of them literally can't be fixed without breaking changes which is why we're in this position to begin with
same with (muli monitor) vrr. x11 is inherently designed in a way that makes this extremely hard. the idea that you can just bolt this feature onto a codebase that's already horrible (according to the people who actually work with it) is insane
Technical_Strike_356@reddit
Your take is the idiotic take. You're saying that nobody can fix X11 without making breaking changes, and the solution is to... make the most breaking change of them all by deprecating the whole protocol and replacing it with a new one with a totally different design philosophy and nothing in common with the old one? Are you hearing yourself?
Infamous_Process_620@reddit
yes
x11 is 40 years old. it's been built with a bunch of core assumptions that are no longer true or relevant. the fact that there are so many of you guys crying on reddit but a grand total of 1 (one) guy who stepped up to actually maintain x11 speaks for itself. nobody wants to touch that shit anymore
arthurno1@reddit
It was 20 years old when they started with Wayland and were parroting the very same argument you allre parroting, already back then. I am watching the same arguments and the same discussion for about 20 years, I think.
Just because things are old does not mean they are bad.
Of course X11 has problems, but it also has some good ideas.
rockymega@reddit
Good, then go fix it. Man. The people working on Wayland now aren't a bunch of idiots, and the people who founded Wayland were X11 developers. They knew that codebase like noone else. They saw what was impossible to build. They saw how many breaking changes they would have had to make from the ground up. That's why they just made a new paradigm using what they had learned. They know X11 better than you. They didn't start with scratch just for fun. They had proper reasons you would get if you were an X11 maintainer. I'm not a fan of all their design decisions, but it can be fixed with less effort.
arthurno1@reddit
People who designed X11 and people who wrote and maintained XFree86 which later become X.org where not at all the same people. What you say would be a valid remark if Wayland was designed by the people who designed X11, or even by people who wrote XFree. Instead the code got into the hands of someone who perhaps never liked it to start with. Who knows. Sometimes there are other driving forces behind people's actions.
At the time, at the end of 90's and early 2000's there were myriad of various "X11 successors". The X11 code was MIT licensed, and was relatively well documented and understood. It was inspiring and easy for people to start building on something similar but with their own visions of how a display server and guis should work.
Some of critique Wayland creator(s) had against XFree and X11 is valid, definitely.
How do you know that? Are you a Wayland developer? Lots of projects start for fun, fame, or profit.
I guess you are an X11 maintainer and you know best :).
arthurno1@reddit
No, it is not. If you think so, you certainly don't understand either X11 nor Wayland.
On the contrary, the separation between the protocol and the library and the design for the extensions from the beginning make it possible to implement new stuff and disable the old one.
For example, there is a valid concern that input can be seen by third-party applications. It is not unlike what we see with http protocol and thay could perhaps be solved similarly with a secure extension to X11 protocol as https or ssh.
A window manager could only see a special input needed to manage windows, and the application input could be encrypted. Each application could see only input they are entitled too.
The other security consideration was X server running ad root, but can be solved similarly by connection from user space via encrypted connection, and of course disable connections from other clients but local host by default. Those who need access via network know what they are doing (hopefully).
kinda_guilty@reddit
The people who built x11 felt/thought/knew adding those would not be possible. The code is free, anyone who feels strongly about it is free to do this. You can't force people to work on something they don't want to in FOSS, that is antithetical to the point of these projects.
slamd64@reddit
Those people who worked at X11 are moving to Wayland. Those hobbist devs who want to contribute to X11 will just see their PRs closed. It is shame that Wayland still is not mature enough after all of these years of development. Worse than that is they are trying to enforce everyone to use Wayland and kill X11 for the sake of change.
I am not against change, but Wayland is just not fully ready yet for production. Those devs are just giving an example where real problem lies.
Generally speaking I feel we are living in bleeding edge world. We are using and buying software that is not tested well and we are QA testers, reporting bugs that shouldn't be in final production release. That is also case for many games.
kansetsupanikku@reddit
You present it as if it was a technical choice based kn merit. While full-time developers are being paid for their work, and the source of that decision is corporate policy of Red Hat and others that followed. "It can't be done" statements about X11 are entirely unconvincing, considering possible implications of the effort that would equal what was given to Wayland. Instead, we get Wayland with its very own list of "it can't be done" scenarios, as described in the post. The truly unfixable one being fragmentation of compositors which make it insanely difficult to support them all. Toy systems with a web browser and some Wine games will work, but GUI apps that need complex workflows are now confirmed to never be getting GNU/Linux ports.
kinda_guilty@reddit
It's open source code. People who love X11 are free to fork it and continue development. Some have actually did recently, though it will take some time for it to be seen if it will be a healthy project in the long term.
kansetsupanikku@reddit
Yes, someone did it because they hate dei, giving a stinky political foundation rather than technical one. Way to go about discouraging participation.
Regardless, you know what? Give me 70% of my current wage and a part on decision making on how I allocate my time to specific tasks, and I will work on X11 full time, any fork you are ready to found. Love won't write the code or support my family. Open source is a way to cooperate and synchronize effort, not to magically create resources such as effort in software development beyond toy projects.
RoryYamm@reddit
The DEI hate came later. He wanted to implement a new X11 extension for security called xnamespaces.
He could not. That's probably why he was so angry about Red Hat and why he thought DEI was out to get him - because the Xorg team, backed by Red Hat, just didn't merge any of his sensible improvements.
kansetsupanikku@reddit
Indeed. The main founder thought that "DEI was out to get him". He might code and code well, but questionable mental state and connection to reality are not what makes a leader trustworthy. And it reflects on the project. Fragile mental state like this could be and will be used by malicious third parties to gain trust, and push code that will get positive review purely on emotional basics.
RoryYamm@reddit
Eh, if you see the current X11Libre development, I don't think that'll be the case.
Oerthling@reddit
It's not for the sake of change.
Maintaining software takes effort. And it's an attack surface that needs to be actively defended.
Replacing X11 with Wayland was always the goal. That's what Wayland was made for.
And you don't have to immediately update to most recent Gnome or KDE (and whatever else project next drops X11 - as more or less all will eventually do). Somebody running, say, Ubuntu 20.24 can use X11 into the 2030s while still getting support. And it's not like Wayland won't get features and fixes in the meantime. To the contrary. Becoming the default and now only option for major DEs means that the remaining pain points get more attention, exactly because people can no longer just easily switch to X11.
minus_minus@reddit
From initial release to X11 in just over three years (1984-1987) ... then X11 for 39 years. Maybe it's time for some breaking changes and X12?
burning_iceman@reddit
If one is willing to make breaking changes, there's no reason to stick with any kind of X. Either you keep enough of X to still be stuck with many of the problems or it's too different to be X in any meaningful way.
Technical_Strike_356@reddit
Of course there is. It's better to make a few breaking changes in X11 than to make the ultimate breaking change by replacing it with an entirely new protocol. Not to mention that X11 has the huge advantage of having one server implementation, X.org, so the changes required on the server-side only need to be made once. Meanwhile on Wayland, any new protocol needs to be implemented separately by every single compositor!
minus_minus@reddit
You don’t have to break everything. The Xorg project lists many functions that are deprecated in favor of more modern solutions. Just dropping those could be a start.
Lux_JoeStar@reddit
How can we make x11 secure in 2025? it's broken from the very core of its foundation. I tried to go through it and fix it but it's a waste oif time, there's no point even trying, it's an impossible task.
slamd64@reddit
X11 likely needs a complete rewrite not just be forked like Xlibre tries to do. I guess X11 code wouldn't conform to today standards of writing clean and sustainable code. It is likely a mess to navigate through architecture.
AyimaPetalFlower@reddit
they couldn't though, it would break apps.
Max-P@reddit
It's important to get right. Where does a window that puts itself at (0, 0) go in VR/AR like an Apple Vision Pro? What even is (0, 0) on 2+ monitors, should that always be topleftmost corner or the corner of the primary display or the corner of the screen the app's window is on?
The solutions that popped off are pretty brilliant and will actually work even in those crazy edge cases. It's futureproofing so we don't need another protocol in 10 years.
Xorg has problems in part because of assumptions like "there's only one screen, one keyboard and one mouse" made in the 70s that don't hold up, so it's worth thinking about those now so it's not a pain in the future.
DoubleDecaff@reddit
Sounds like someone wants to help out programming features.
fellipec@reddit
One day Wayland will be great.
Oerthling@reddit
For plenty of people it works totally fine right now.
somethingrelevant@reddit
big line between "works on my machine" and great software
Oerthling@reddit
Sure. But also big line between still has some problems in corner cases and bad software.
coyote_of_the_month@reddit
I've been using Wayland daily for years. How many years? How should I know? One day it became the default in Gnome and it just worked.
roberp81@reddit
only for people that uses their pc for Facebook
roberp81@reddit
like 20 years maybe
zlice0@reddit
holy shit...no way. warping? i thought that was never going to happen
involution@reddit
engage
omniuni@reddit
Great, now we just have to wait for at least 3 window managers to implement it.
Wonderful_Turnip8556@reddit
SDL already implemented it
TiZ_EX1@reddit
That only means that applications using SDL will attempt to use the protocol. The window manager side still has to support it.
fiery_prometheus@reddit
I hope this means that Wayland finally choose practicality in some areas, over rigid security concerns, and this is a step forward to getting a more unified API, with the security being taken care of in another way which doesn't put in a stick in the whole goddamn ecosystem of window managers on Linux.
RanidSpace@reddit
now im interested, a lot of games still work on wayland beinf able to lock the cursor so it doesnt move away, how was that dealt with?
OskaroS500@reddit
Could someone explain what is the difference between X11 and Wayland? I know that X11 came before. What was wrong with X11 that people created something new?
FriedHoen2@reddit (OP)
Few things, surely emendable. In 2000s Xorg was heavily modernized by very capable devs, mainly from Intel. The today Xorg uses the same technologies that Wayland uses (Direct Rendering via the kernel drivers for example).
After that, young devs from Red Hat were unable to modernize Xorg so decided to pick a project born for embedded devices (Wayland) and adapt it. They also decided that Wayland is only a strandard, not a program, so each desktop environment has its Wayland implementation, making fragmentation a huge problem in the process.
It took more than a decade to obtain something barely usable. Neverthelless they decided to kill Xorg and force users to switch to Wayland, while it lacks (or has bad implementstions of) some basic functions for modern desktop.
Informal-Clock@reddit
As someone who writes and reads Wayland core everyday a lot of this just sounds like excuses. Clipboard is possible as is and pointer warping is already possible even without the protocol! Additionally, the issue with throttling is a client problem afaik there are ways to deal with/detect it so, not the driver/underlying system fault. Positing dialogs has been possible for eons, only thing you can't position are top-level surfaces which is understandable if they are being limited by that
NECooley@reddit
Yep, Wayland sorta sucks. Unfortunately, x11 is worse.
RoryYamm@reddit
Better the devil you know…
blackcain@reddit
Well, they can show up at Akademy and at GUADEC to talk about these things. You have to show up and participate. Writing blog posts is fine in getting the word out but somenoe from KiCad should show up.
We have an entire freaking conference related to applications that I help organize called Linux App Summit. Maybe show up?
RoryYamm@reddit
Why would they bother, when you've been going at it for 18 years and the only response to 'hey, why can't I port this application with this UX that's worked forever' seems to be 'well, maybe you should rework your UX design then sweaty!'
Christ, at this point in X's life, it was running on practically every OS, including Windows - and was THE way to get graphical applications streamed to remote machines. Hell, in 5 years time, the Wayland project will be as old as X11 was when it was deemed 'unmaintainable' and focus shifted to Wayland!
blackcain@reddit
I was around for X10 and X11 releases. The number of times I would get a root prompt randomly on an X windows workstation... (it used to happen on NeXT workstations too, lol)
You all worshipping this stuff like it will never age and is perfect forever is puzzling at best.
Since I lived the lifetime of seeing X start and end, I've moved on. Sometimes, you just need to start fresh. In about 5 years time, you all will forget all about X windows other than a piece of computing nostalgia like we talked about Vaxs and IBM system/370s.
RoryYamm@reddit
Well, my apologies that you got to experience the bad old days of the X Server. Problem is, a lot of us weren't even born back then. We instead enjoy an extremely stable X11, where we can use the applications that have worked for our entire lives.
Now, the unstable graphical server is Wayland. However, unlike in your time, when you had very little choice, we can go back to X11 - at least, until you rip it out of the popular desktops and widget toolkits in the name of 'progress' sometime in the next year or two. Oh boy, can't wait to be forced to use something that frequently screws up and never quite works right if we want to use any halfway popular application!
You've been developing this stuff for the better part of two decades. Two decades that could have gone into designing an almost-compatible next-gen X12 server or finally addressing all the technical debt of X11, instead went into a gigantic bikeshedding exercise, only dragged kicking and screaming to something half-functional by Valve. In the process, you've managed to create half a dozen incompatible implementations of your precious 'protocol', and a baffling security-theatre system that renders annoyingly simple and useful things impossible without another protocol extension that's always coming just next year. Hell, I KNOW it didn't have to be this way - were it not for your team at GNOME stonewalling every sensible idea that didn't fit your Jovian view of the desktop, I might actually be running Wayland at this point, happliy saying goodbye to WindowMaker.
At this point, shit or get off the pot.
X_m7@reddit
Yeah, this damn attitude is the thing that makes me dislike Wayland and GNOME, I've seen way too many issue reports getting whacked with such sentiments, and hell even in pull requests where someone else did the work already so it's not even a question of "we don't have time to work on it so we're looking for reasons to ignore it" either, ugh.
FriedHoen2@reddit (OP)
Sorry, why should they? They have made it clear that Wayland's problems must be solved upstream. Some are not solvable by choice. So there is nothing to discuss basically.
blackcain@reddit
because ultimately it's an app that is running on their desktop. It's how you advocate for your project. I mean even GNOME apps might want pointer warping, right?
A lot of upstream wayland people are also GNOME and KDE people.
FriedHoen2@reddit (OP)
I still don't understand. KiCAD developers and its users will be able to continue to use X11 at least until 2032 when RHEL 9 will run out of support and then Xorg will probably be abandoned completely. Until then perhaps Wayland will somehow correct the problems encountered, and if that doesn't happen, KiCAD will continue to run fine on other operating systems. They don't seem to be in any hurry. If XLibre succeeds, it may even go beyond 2032. We shall see.
blackcain@reddit
XLibre is not going to succeed. The man is a crackpot and his code is filled with issues as noted by the actual maintainers. Feel free to think whatever.
Eventually, the drivers under Xorg won't be as well tested. There are no more releases of Xorg under than critical security bugs.
But once the switch happens then Wayland development will move faster as companies and others switch over. In the meanwhile, the developers can engage with the app ecosystem.
Nobody expects app developers to be involved in Wayland upstream development, but they can talk to the people who manage their toolkits under Linux.
FriedHoen2@reddit (OP)
Maybe. Wait and see.
Vindve@reddit
Something I don't get about Wayland is why every window manager or compositor has to implement its own server instead of ensuring there is a default implementation of the protocol that is the default and used by everybody, like XOrg was. Seems a lot of wasted effort by everybody.
imbev@reddit
It's very easy for a "default implementation of the protocol" to become the protocol itself.
Vindve@reddit
I (highly) understand this problem for some protocols like network protocols where you really want to have a competitive environment, but here, it doesn’t seem to me a problem if there is a "main" implementation of the protocol on which any desktop environnement can plug instead of reinventing the wheel. We want to have a competitive environment for desktop environments, but for the display server, I honestly don’t care if there is one that overtakes any others in the Linux world.
RoryYamm@reddit
Why is that a problem? If that one implementation of the protocol is well-documented, and everyone can actually target it, thus allowing it to always work, how is that a bad thing?
Right now, a Wayland application either has to build for a dozen different implementations or stick to being KDE- or GNOME-only. How is that an upgrade over X11? At least X11, for all its flaws, was going to be the same flawed X11 no matter where you found it. You could anticipate it, you could deal with it, you could rely on literally everyone else being in the same boat for help. Now? GNOME solutions don't work on KDE, KDE solutions don't work on GNOME, and let's not even get into Hyprland, who said 'fuck this' to all that bullshit and went off to do their own thing.
imbev@reddit
It's not inherently a problem, but it can cause problems.
We can see the dangers of an implementation-defined standard by looking at the history of the web. If your browser isn't Internet Explorer or Chrome, it might be discriminated against.
I don't think that anyone would have an issue with a monoculture of X11 if it supported everyone's use case. Wayland was specifically designed to be the "perfect" solution, one that is flexible enough to meet everyone's use case. Which as you rightly point out, has caused Wayland to have slow development and inconsistent experiences across platforms.
RoryYamm@reddit
One of the criticisms of X11 is that it's too bloated, trying to do everything at once, trying to be everything… but do go on about how it doesn't support everyone's use case. Seems to support far more than Wayland at the moment.
I mean, for crying out loud, Wayland is 18 years old at this point. X11 at 18 had just barely started its SECOND fork for x86 machines (X386 became XFree86 became XOrg), and was the gold standard for windowing systems other than Windows. Hell, even WINDOWS had X11 support. Wayland now barely has support for Linux, with all the graphical problems people still experience on KDE and GNOME.
InfiniteSheepherder1@reddit
Some initial prototyping happened but main Wayland development didn't start until the 2010s. If you are going to count from first description of the project then you should start with X in general, rather then X11
Wayland is closer to 12 or so for when development really began. Fedora moved to Wayland by default on 2016, 3 years after GNOME worked on moving to it. I have been running it mostly full time since then, and in 2018 was the last time I loaded into X on my desktop. So to me Wayland was usable full time within 4-5 years
I wouldn't say gold standard, DRI had to be implemented to do 3D graphics and was not until 2009/) windows could get their own buffer to draw to instead of a shared one with DRI-2 developed by the author of Wayland the same year he initially proposed it. There was other limitations and all this is why.
Even in the early 2010s we didn't have accelerated web video in part due to issues with X11 and getting things without tearing especially when moving windows was pretty bad.
taicy5623@reddit
Didn't mouse warp JUST get merged by KWIN?
Krunch007@reddit
Just more blabber. Use it through XWayland, you don't have to use it natively via Wayland if they can't pour more support into it, which is understandable. Still, more of an absolute nothingburger.
mort96@reddit
I don't think XWayland applications can do things Wayland applications can't do, so things like wrapping the cursor and positioning windows aren't solved by just using XWayland.
Krunch007@reddit
Yeah it can, considering it's running an xorg server just for a Wayland window. You can do cursor warping inside XWayland just fine as long as you're warping it inside the window. Of this I'm sure because multiple other apps can do it and I've also done it myself. Window positioning I think is also possible like that, but I'm less certain because I can't recall using an app that uses those at all...
_chococat_@reddit
Kicad uses different windows for different views of the circuit. Warping within a single window isn't their use case and isn't useful in the usage patterns of the application.
SuAlfons@reddit
looks like it's time for Kicad to think about their user interface.
RAMChYLD@reddit
Agreed. XWayland is a thing. Just because you’re running Wayland doesn’t mean X programs will drop dead.
Pabloggxd123@reddit
to report an issue, you must confirm that it happens with x11, if you say that your running xwayland then they dont care.
Kicad is an amazing project, nothing against it.
FriedHoen2@reddit (OP)
I also omitted "we dont accept Wayland-specific bug reports". Translated: we dont give a fuck.
alexforencich@reddit
It's called prioritizing. They don't have the resources to rewrite the entire application to work around stuff that Wayland is missing.
mrtruthiness@reddit
There's a difference between "not accepting" and "prioritizing". They could accept it with low/no priority, right???
alexforencich@reddit
I think that's possibly what they mean... Go ahead and open it, but we'll probably just ignore it.
mrtruthiness@reddit
I don't use Kicad.
Frankly, the post seemed to be "drama seeking" which I find horrible. Furthermore the post was very unclear about which aspects of Kicad were not workable under Xwayland.
mistahspecs@reddit
Please don't become a translator, you're quite bad at it
LowOwl4312@reddit
Bug reports are not pull requests. It would be just a load of bugs closed as "WONTFIX upstream issue" anyway.
Krunch007@reddit
Very clearly they care enough to direct devs towards Wayland development, so I'm not exactly interested in your inferred translations.
anotheruser323@reddit
Yep. ICCCM and EWMH.
I said it from the start, but everybody's like "oh, wayland is the future". This is really basic stuff that is missing for 16 years.
devonnull@reddit
It's the applications and users that matter, not the display protocol. Wish the waylanders understood that.
CoronaMcFarm@reddit
X11 is not even maintained, what is the alternative?
crazy_penguin86@reddit
It actually is being maintained still. I don't know why this gets spread around.
Infamous_Process_620@reddit
what does maintained mean here exactly? like security issues get patched (the ones that aren't inherent to the protocol...) but there is no development AT ALL in any 'new' features
crazy_penguin86@reddit
That's... being maintained. When you maintain a bike, or a lawnmower, do you go in and add new features to it? The answer for almost everybody is no. You just make sure it keeps working. A lot of people here seem to want to stretch "maintained" to also mean "adding new features".
SEI_JAKU@reddit
It's the same kind of brainrot you see in games. If a game isn't constantly being updated with "new features", the game is dying/dead. Doesn't matter what kind of game it is.
gib_me_gold@reddit
That's called feature completion.
Infamous_Process_620@reddit
so x11 has been around for 40 years and is still not feature complete? and people complain about wayland being slow lmao
gib_me_gold@reddit
There is no new developement because no new features are needed. That's what I said. Reading comprehension is not your strong suit, eh?
Infamous_Process_620@reddit
i want and need those features. what now fucko?
Kevin_Kofler@reddit
There is a fork that is maintained.
kinda_guilty@reddit
It's barely two weeks old, "maintained" is an exaggeration.
Makoto29@reddit
I can't believe someone forked it. wtf?
VividGiraffe@reddit
I think because the user submitted tons and tons of PR fixes and they were all shutdown. So he instead forked it. Meh
Huehnchen_Gott@reddit
As it seems right now, nothing
natermer@reddit
Yeah it is too bad that the guys that write and maintain Xorg doesn't understand that.
devonnull@reddit
That's kind of my point. Xorg just works.
natermer@reddit
Xorg and Wayland "dummies" are the same people.
devonnull@reddit
Maybe they should just stick with one that has better documentation and doesn't pander to fringe cases of use.
Krunch007@reddit
You really don't see the irony in writing this considering your cultish behavior? You can literally use KiCaid on your systemdless xorg system without a hitch. What are you complaining about, even? That people don't wanna bust their balls maintaining shit that doesn't work in this day and age? Go develop it then.
Want me to throw in a hundred anecdotes of xorg graphic glitches, color infidelity, input issues, crashes, etc, basically xorg not working where Wayland does? I could. Does that make either of our points better? I don't think so. I think you'd be better served by not complaining about not being understood and instead signing up to maintain xorg or XLibre and use whatever you like.
devonnull@reddit
Cultish...whatever...with your response you need to look in the mirror. I remember when Wayland was going to be the default display for Ubuntu 16.04 or whatever version the 'real soon now' crowd was pushing at the time. I'm sure in 5 years some one is going to pull the same argument of old crusty code blah blah blah and start a new project to do away with Wayland...systemd....etc...
It just feels like all the software is now is virtualized wrappers which is a troubleshooting mess (oss/alsa/jackd/poeterringshitware/pipewire).
grem75@reddit
I remember Shuttleworth saying 16.04 was going to be Mir by default, certainly don't remember them ever saying they'd be on the bleeding edge of Wayland. They never said there was going to be an LTS with Wayland as the default until 22.04, they walked back the Nvidia side of it but that did happen.
Fedora was the one to watch for Wayland adoption. They were Wayland by default in version 25 in late 2016 for Intel and AMD users, which was pretty close to their prediction.
Of course RebeccaBlackOS was shipping Wayland by default in 2012.
devonnull@reddit
Possibly, maybe.
Mir, Wayland it's all the same to me at this point. Enshittification of Linux.
iamthecancer420@reddit
I can understand Wayland because a lot of legacy programs break but systemd literally just werks and is 10x simpler for the layman (who shouldn't even care about PID1) over SysVInit and has been the default for over a decade now.
devonnull@reddit
Just wait until they make the unit files binary.
Kurgan_IT@reddit
They don't, they truly believe that their way is always the right way and everyone else is an idiot.
devonnull@reddit
And they're the first to throw around terms like cult.
Puzzled-Guidance-446@reddit
It's fun how there is a lot of this on linux communities generally speaking.
kalzEOS@reddit
It was already very hard to get companies to bring their apps to Linux.
Linux now: "here is Wayland. Now get lost forever".
ECrispy@reddit
Wayland fixes a lot of X11 cruft, but these points are valid, its not really well designed or thought out, its a half baked set of protocol specs that basically shifts the burden to implementers and doesnt provide any standardized benefits.
ABotelho23@reddit
Wayland is a protocol like X11 is.
People seem to forget that there used to be many X11 servers. Eventually all fizzled out except XOrg.
That might happen again. There are similar projects, like wlroots: https://github.com/swaywm/wlroots
Zettinator@reddit
Yeah, it was a major oversight to just focus on the protocol only. A reference implementation - or maybe two different ones for different use cases - would have been just as important.
nightblackdragon@reddit
It wasn't oversight, it was on purpose as the fact that there is de facto one implementation of X11 is not that great for everyone.
Zettinator@reddit
Yes, but that doesn't mean that the opposite is a good idea. It is bad to only have one implementation (that essentially ends up defining the spec), but having 5 or so is a waste of resources, too.
dr_eva17s@reddit
There is a reference implementation called Weston.
Zettinator@reddit
Weston is designed more like a "toy implementation" used to experiment with the specification. It never was intended as a practically usable display server.
AyimaPetalFlower@reddit
there's too many good implementations of wayland display servers now. There's also smithay and mutter, I think mutter is made with "libmutter" or something so theoretically it also has its own library but I'm not sure.
It's not nearly as hard to support wayland protocols than it would be to create a new x11 server and it probably won't ever be.
YellowOnion@reddit
You don't say "good implementation" and mutter in the same sentence, it has terrible performance, and terrible design, quite literally can't implement server side decorations without a massive rewrite, which is supported by all the other major compositors, even Cosmic, which is only 3 years old.
AyimaPetalFlower@reddit
I don't know if that's actually true and I have decided it doesn't matter
mishrashutosh@reddit
wlroots will never be the primary implementation as long as GNOME and KDE are doing their own thing. I think COSMIC also has its own Wayland implementation? It's too fragmented.
YellowOnion@reddit
It's not really that fragmented, the problem is Gnome devs stonewalling the commitee process, and the committee process in generally encouraging bike shedding instead of fast iteration.
Generally the fragmentation is just "gnome" and the others, KDE tends to be one of the bigger contributers to experimental features, and wlroots/sway also does a decent job at implementing new features.
And if there's some fragmentation between the compositors, it's because it's still an experimental feature, or lack of dev work needed, And I would rather have a bunch of fragmentation around experimental features to figure out what users actually need before a standard is standardized. Valve attempted to speed this up with frog-protocols, but the freedesktop guys took that as a wakeup call to adjust their staging process to reduce bike shedding.
Aln76467@reddit
don't forget smithay and aquamarine
ztwizzle@reddit
This will never happen for Wayland because of how the protocol standardization process works. If GNOME/COSMIC/KDE/etc gave up their Wayland implementation and moved to wlroots, they would lose the ability to vote on protocols. I'm pretty sure this was by design, so people aren't able to get away with underspecifying behavior and having the real spec be "whatever wlroots does" like what ended up happening with X11 and Xorg.
Zettinator@reddit
Still, KiCAD's approach of not even accepting bug reports specific to Wayland is completely idiotic. You can't improve Wayland support that way. And it is inevitable that they need to support Wayland officially, very soon actually.
londons_explorer@reddit
"we don't accept bug reports, but we will accept pull requests".
Basically, we don't have any intention of fixing these problems ourselves, so please don't put effort into reporting them, but if you want to fix it yourself then that's fine.
Fit_Flower_8982@reddit
If there are no bug reports, third parties will not be able to submit pull requests for bugs they don't have themselves, nor take them into consideration. Better to just tag and ignore.
londons_explorer@reddit
It is super rare for third party 'drive by' contributors to fix some bug they don't have themselves.
PureTryOut@reddit
That's strange though. Having a bug be reported does not at all imply only the maintainers can fix it. It's just documenting of an issue, that's all it is. If someone wants to fix Wayland problems it can be useful to have a list of things broken on it, but if you're not accepting bug reports for it there will never be such a list.
Altruistic_Cake6517@reddit
Which is entirely reasonable, tbh.
They have to prioritise their time. It's a bit of a gamble that it'll "work itself out" in time though.
RoryYamm@reddit
I doubt they'll have to. The people that use graphical Linux to make money don't need GNOME or KDE - they need KiCad, or whatever other X11 program they've been using since CDE and AIX. X11 is also the only game in town on BSD.
snippins1987@reddit
That's not idiotic, that's pragmatic and saving resources for what matters. I'll continue to treat Wayland as something that does not exist, wait for it to finally work and switch to it, I'm perfectly fine with Wayland full adoption take another 10 years because there is nothing from Wayland that makes me really excited at all. It's especially not fun looking years long discussions about a feature you need. I spent years perfecting my Linux workflows and unless I can replicate them 1:1 in some Wayland compositor, then I won't touch Wayland.
Talking about what's idiotic, I think it is better to describe development process of Wayland. Everything is so slow to progress, so many usage problems for both developers and users, and being pushed everywhere even though the protocol itself cannot cover plenty use cases.
Yes, someday I'll switch because eventually people will finally solve Wayland, once it becomes something good - either by itself or by all the workarounds people made for it, though I bet it's the later.
Godbless the people who will put in the work to workaround wayland to support all the insane workflows and keeping Linux interesting for the future tinkerers.
gib_me_gold@reddit
It's not their fault though? Why would they accept bug reports for something they cannot support?
Also
>And it is inevitable that they need to support Wayland officially, very soon actually.
kekw
feldoneq2wire@reddit
Feature requests sit in the KiCad queue with without being addressed all the time. A Wayland megathread would be completely appropriate with the clear explanation that it's not on the KiCad dev radar. How can you do pull requests if there's nothing to point at and say "this is what I'm addressing"?
FriedHoen2@reddit (OP)
And it is inevitable that they need to support Wayland officially, very soon actually.
Why? Xorg will be supported until at least 2032 since the EOL of RHEL is 2032. It doesn't seem to me that the kicad developers have anything to worry about for the next 7 years. Maybe Wayland will be working properly by then.
ECrispy@reddit
and when they do I bet they will accept those bug reports, till then they are just noise and serve no purpose.
people forget these projects dont have resources or people to look at everything.
roberp81@reddit
but wayland add a lot... really a lot of problems that X11 doesn't have
sizz@reddit
Wayland is extremely hostile to disabled users by design. How is widespread adoption going to happen when by law, government and companies need to set up a computer accommodating disabled people.
ECrispy@reddit
its happening already, because companies like RH with their $$$$ and clout have defacto control over Linux, and Linus, the only one who can do anythying about it, doesn't really care about anything besides the kernel. The day he goes that also will turn to shit and be controlled by corps.
Its funny how Windows and MacOS are far more disabled user friendly, have actual usability guidelines that are followed, but still nothing close to whats needed, esp by new apps.
FattyDrake@reddit
Apple can and does force apps off their platforms if they don't follow basic accessibility guidelines. If your app breaks when someone uses an accessibility setting, probably will be rejected.
Microsoft controls some of this through their app store as well as whether an app can officially say it works with Windows (using logos in marketing, etc.)
Which authority on Linux can reject apps that do not follow basic accessibility guidelines?
The closest might be the desktop environments. Can you imagine the uproar if GNOME said "Any app that doesn't follow guidelines will be blocked from running on GNOME."
This would be a level above KiCad just not wanting to update for Wayland.
ECrispy@reddit
this happesn although not for the reasons you think. Why do you think other distros fork Gnome all the time?
GrayPsyche@reddit
Valid and deserved criticism, but saying x11 is better is just not true. x11 is a security nightmare. It's also not very optimized and carries a lot of old baggage. It's unusable for me.
FriedHoen2@reddit (OP)
Why would Wayland be better? On the contrary, placing security at the level of the GUI only gives you an illusion. That is not the right level to isolate applications.
I don't think so, all tests say that Xorg and Wayland are roughly equivalent in terms of performance.
gmes78@reddit
That is a stupid way to view things. Having a secure display protocol doesn't magically make the whole system secure, but if you want a secure system, you need a secure display protocol. It's just one piece of the puzzle.
FriedHoen2@reddit (OP)
Even if this were true, which it is not, there is no point in boarding up the windows and keeping the door open.
gmes78@reddit
The Wayland developers are only responsible for the security of their own project, not the security of the entire system. Is this so hard to understand?
FriedHoen2@reddit (OP)
Then Wayland fans should not 'sell' Wayland as the solution to Linux security problems, and especially should not delude themselves that it is.
No, because it is fundamentally wrong. You have to solve security problems at the kernel level and other basic libraries, once you've done that, then you won't need security at the graphical level, or in any case it will only be a secondary level for very specific cases or to warn the user.
gmes78@reddit
No one does. You're dreaming up a strawman.
That just shows you don't understand anything. A protocol such as Wayland IS ONE OF THOSE BASE COMPONENTS. You CANNOT change it after the fact, you have to get it right the first time.
FriedHoen2@reddit (OP)
On the contrary, defence in depth is exactly what I was saying. For instance, isolating input between processes is not something that needs to be implemented at the graphics server level, but at the kernel level or at the level of basic system libraries. In any case, the security holes in X could also be fixed in X Nothing and nobody prevented it.
gmes78@reddit
You're just proving my point: you don't understand anything.
Access to inputs is already restricted at the system level. It's why ydotool typically needs to be run as root. (Wayland servers get handed control of those devices through a seat manager, such as seatd or systemd's logind.)
The problem was never that the kernel or base system were doing something insecure, it's that X.org itself gives everyone access to inputs.
Also, you conveniently forgot about everything else that's also security related, such as X11 allowing any app to view the contents of other apps, or the entire screen, without needing any sort of permission. You cannot fix a security hole like that from any other place on the stack, as it's caused directly by X11.
FriedHoen2@reddit (OP)
The apps involved would be exactly the same as those giving problems with Wayland and the solution would be the same, give them permissions.
This is only partly true, in fact, you can write a Wayland keylogger working without root permissions.
The exact same problem is present in Wayland because of the way the input management is designed (in both Xorg and Wayland). The difference is that Wayland puts a patch on it, which could just as easily have been done in Xorg.
gmes78@reddit
I'm done replying to your nonsense.
FriedHoen2@reddit (OP)
Or maybe you haven't any valid reply.
gmes78@reddit
I'm sorry, but this is one of the most one-sided comment threads I've seen in a while, with how obvious it is that you're wrong. Anyone with half a brain can tell you that.
FriedHoen2@reddit (OP)
Ok, man with 1/4 brain.
sizz@reddit
For decades I heard Linux is more secure than windows, now x11 is a security nightmare. Is there an example of a real life hackers penetrating Linux systems via x11?
Accurate-Sundae1744@reddit
And then one wonders why companies just makes electron apps.
pppjurac@reddit
Or simply flat out refuse to invest funding to port application to linux desktop.
FriedHoen2@reddit (OP)
Nobody will invest in Desktop Linux having to support this fragmentation.
xplosm@reddit
They do it for Blender, GIMP, Krita, OpenCAD…
nightblackdragon@reddit
They didn't want to invest funding to port application to Linux when X11 was the only option.
Negirno@reddit
And why wouldn't they?
Most applications on mobile are basically just a webpage running in their own sandbox.
Preisschild@reddit
Electron supports wayland and id rather use an electron app than some more native frontend that has access to all my windows and input devices...
UgglanBOB@reddit
Are you a friend of the nazi that forked X11?
Technical_Strike_356@reddit
Nazi?
Kevin_Kofler@reddit
berarma@reddit
Wayland is here to remove all the X11 issues we didn't get to experience and bring new issues we're already experiencing.
Weird_Tomatillo1323@reddit
Are fair issues.
Arguably bad UI paradigms or don't really affect functionalities.
Most likely their own fault.
FriedHoen2@reddit (OP)
As usual, someone else is always to blame.
If it worked fine under X11 and the problem only occurs on Wayland, Wayland is to blame, not the app.
nightblackdragon@reddit
>If it worked fine under X11 and the problem only occurs on Wayland, Wayland is to blame, not the app.
It doesn't work that way. Wayland is not supposed to be compatible with X11 so you can't expect X11 app to run flawlessly on Wayland. Are you also blaming Linux for not running Microsoft Office? it works fine on Windows.
FriedHoen2@reddit (OP)
Exactly. Kicad works fine on Win, Mac and X11. Not fine on Wayland. Do the problem is Wayland.
T0astedGamer03@reddit
I mean they said nothing wrong. They said a few of the issues are a fair reason that the app isn't Wayland friendly yet. Then they said that something that doesn't affect app functionality isn't a real showstopper (which yes an app remembering where it was previously launched isn't a required feature for the functionality of a CAD app) and that the performance problems are probably the fault of the codebase and not Wayland itself, especially with QT and KDE libraries working well with Wayland.
FriedHoen2@reddit (OP)
You seem to have misunderstood the context, those shortcomings make using the app uncomfortable and they themselves say they only use it under Xorg.
JackDostoevsky@reddit
i think the fragmentation issues are a bigger issue than window positions or cursor warping: those things exist in various compositors in various ways
it's the fragmentation mentioned that's the biggest issue. ironically i feel like targeting wlroots protocols is probably the best idea: GNOME doesn't support them (at least last i checked) but Plasma does.
not really providing that as a solution, more just as a comment on the silly state of Wayland development: wlroots exists because of the shortcomings identified in this post
FriedHoen2@reddit (OP)
kwin is not based on wlroots. There is a fork of kwin that is, theseus-ship
JackDostoevsky@reddit
no, it's not, and i apologize if i implied that (i don't believe i did). kwin does support several wlroots protocols, though.
Potential_Penalty_31@reddit
then xwayland? I don't get all the drama
Pabloggxd123@reddit
if you have an issue, you must confirm that it happens on x11 session, if you run kicad with xwayland, they wont care
FriedHoen2@reddit (OP)
Xwayland dont solve most of problems listed in the post.
arades@reddit
How does it not? They keep shipping x only application, people with Wayland will just run it through xwayland transparently
FriedHoen2@reddit (OP)
Please red the post by kicad devs. Their problems cant be solved by Xwayland.
mrtruthiness@reddit
Their post does not mention Xwayland at all.
ZunoJ@reddit
But when I say this in a discussion about wayland the fanboys won't stop arguing about it lol
creeper6530@reddit
Sums up the entire article great
spectrumero@reddit
I'm really underwhelmed with Wayland. I badly would like something fresh and modern to replace the rather creaking old X11 display server; Debian now does Wayland by default but within a couple of months of using Wayland I always find some irritation or something that flat out doesn't work which forces me to switch back to X11.
Lightprod@reddit
You're running a 2 year old version of Wayland.
gmes78@reddit
I'm pretty sure your issues come from the geriatric package versions that Debian ships, and not due to the current state of Wayland.
FriedHoen2@reddit (OP)
Same here. I try Wayland on KDE with each new release of Plasma. For goodness sake, the improvements are there, but there is always something (badly) wrong with the applications I use the most.
Inside_Jolly@reddit
No idea what Wayland's problem with cursor warping, but (IIRC) both ICCCM and EWMH postulate that an application can position itself but a WM is always free to override that positioning. So, you can't rely on X11 honoring application's positioning either.
FriedHoen2@reddit (OP)
it is obvious that WM should be able to impose a position for special needs, this does not imply that apps are forbidden to have the possibility to position themselves.
PracticalResources@reddit
I hope that guy forking x11 finds success.
FriedHoen2@reddit (OP)
Me too
raven2cz@reddit
Unfortunately, that’s a wish I’ve always had too — but it’s a battle that’s already lost. Most desktop environments have realized that as well. The only real option now is to get Wayland into a more mature and usable state — even if that means accepting some radical changes along the way.
There are quite a few fundamental things I personally disagree with in Wayland, and they’re also part of the reason why its development has been so slow and dragged out over more than a decade. But sadly, there’s no other path forward anymore.
Kevin_Kofler@reddit
Why is it a battle that is already lost? X11 still exists, there are people who want to take it off "maintenance mode" back into active development, and I do not see why we should accept being forced into a purported replacement with less functionality. This is not some crappy proprietary "take it or leave it" platform, this is Free Software and forks exist.
Business_Reindeer910@reddit
because KDE 7 won't support x11 anymore (sometime in the next 5 years probably), GNOME won't in about a year. The new COSMIC is wayland only. All these new wayland only "WMs" don't.
It is going to die via that.
Kevin_Kofler@reddit
Everything that drops X11 support will be forked.
ronaldtrip@reddit
Who will maintain and develop the forks? Not saying it is impossible to do so, but where are these developers chomping at the bit to keep the X11 back end dependent software going? Realistically, it means doubling the current crop of FOSS developers to be able to have two branches.
Business_Reindeer910@reddit
sorry, but that's just not true. it's not going to happen.
gmes78@reddit
And the forks will see no meaningful activity and be abandoned in a few months.
AyimaPetalFlower@reddit
It's obvious why xlibre doesn't work if you have proper context, there's a thing called "rootful xwayland" and anyone could create a wayland display server that supports x11-y things and leaving out the parts that don't fit their vision for the future of the x11 protocol.
This lets you avoid the problems with DDX xorg and full compatibility with x11 and the choice to support wayland only apps and even wayland apps that use custom protocols you'd like added.
You can also use the legacy drm/kms api and do literally anything weird the xorg server did I'm pretty sure too.
Basically it makes no sense to make a fork of the xorg server when xwayland would be way easier to extend than fixing the architectural issues with xorg
S7relok@reddit
So to use this software you need a dependency software that's already condemned
FriedHoen2@reddit (OP)
Xorg will be supported at least until 2032 since the EOL of RHEL 9 is 2032. There is enough time for Wayland to mature.
londons_explorer@reddit
Just ship kicad in a little VM with X11...
blobjim@reddit
A lot of those features end up being clumsily used by applications and lead to horrible user experiences. Having applications place windows and mouse cursors has led to all sorts of horribly designed Windows applications. So hopefully Wayland replacements for them are more restricted.
Using multiple little breakout windows for a single applications has always been a horrible user experience and is something that should be moved away from anyways (although this is something supported by Wayland which XWayland allows KiCAD to use, just not positioning the windows I guess).
And the main window position *is* restored in XWayland/Wayland, but what they want is the ability to restore a bunch of little windows to specific positions, presumably relative to each-other. Which is so much complexity for such an anti-pattern of application usage.
FriedHoen2@reddit (OP)
You want to impose your view to kicad devs. A display server should be neutral.
Infamous_Process_620@reddit
"A display server should be neutral"
literally no reason to believe this
arades@reddit
It makes sense that a windowing system/display server would support patterns users enjoy first
FriedHoen2@reddit (OP)
It should support multiple patterns to be neutral and general. Windows do it, macOs do it, X11 do it.
AyimaPetalFlower@reddit
macos's display server is much more restrictive than you might think
arades@reddit
Yeah, and the protocols they want are underway, but they're rarely used by devs and users don't like them so they haven't been finished yet. If KiCad wants support they kind of need to drive it. You can say that because other platforms support it that Wayland should, but color profiles / HDR support are standard on MacOS, Windows and Wayland but will never be supported on X. It's all about priorities, and those priorities are driven by either developers or users.
k-phi@reddit
This reminds me how people tried to use outdated browsers to continue to use flash player
tesfabpel@reddit
We already discussed this. Also, most of the problem they have are not because of Wayland but because wxWidgets hasn't yet improved its Wayland support.
You can see other professional apps that works fine under Wayland. And yes, pointer warping (not pointer locking which is already there) is being released right now.
MadeInASnap@reddit
Wasn't this kinda chosen for security reasons, or do I misremember? E.g. one problem with X is that SSHing with X Forwarding into a compromised machine lets the remote machine see everything on your desktop. I.e. if you SSH into the remote, the remote can spy on you.
VVine6@reddit
there are PRs to wayland protocols being worked on that tackle exactly these issues. how is this "ommited by design"?
alexforencich@reddit
Well if they're being worked on now, then they must have been omitted from the original design.
gmes78@reddit
Pretty much everything was omitted from the "original design". On purpose. The core Wayland protocol doesn't even have windows!
Protocol extensions are how Wayland was designed to be built.
InfiniteSheepherder1@reddit
I can promise you none of what we do today was possible on 1980s X either. DRI which is what permitted usesapce applications to use the GPU, without it none of this would be possible. It is impossible to know everything that could possibly be needed, but also letting clients do whatever was not an option. Devs worked on what they felt was more important. I kind of agree limiting this stuff to prevent annoying applications is nice.
alexforencich@reddit
It seems like there's sort of a long term evolution taking place of compartmentalization and isolation, for various reasons (not just with Wayland, but also things like immutable distros, containers, snaps, etc.). I get the idea from a security perspective, but it's really hard to graft something like that on to existing systems without having to make compromises. Either existing apps have to be reworked (which can be a big ask for open source projects with limited resources), some features simply won't be possible (explicitly by design, but some people might not be ok with giving that sort of thing up), or the isolation will have to be broken down somewhat. What works for one application might be a serious problem for a different one. Like many things, the truth likely lies somewhere in the middle, but you can get some serious zealots on both sides who insist they're right and no compromise is possible. Wayland explicitly started out with a small footprint and limited set of features, but I think they've finally come to realize that was a bad idea as it has resulted in a lot of fragmentation that's caused a lot more work for application developers to deal more directly with the idiosyncracies of each DE. I think this was partially to favor isolation and partially to try to keep things simple. I guess we'll see how it pans out long term as Wayland finally implements some stuff that probably should have been standardized earlier.
InfiniteSheepherder1@reddit
A lot of that is because we are starting to take security seriously, Microsoft is dealing with this too, they are trying to phase out NTLM which uses MD4 which as of right now by default is still used on Server 2025 in a Domain. Microsoft told people it was deprecated and move off in 2004, we still get software in 2025 at work that don't work. Supporting Kerberos or other tech does take work.
We are moving from giving applications roughly access to anything your user can to more limited, things like selinux and chroot existed and firejail too, so the idea is not fully new. I was setting up Linux at a public library in 2010 and we used firejail for the browser, it worked well but it was pretty hacky. I can get most of the features of that setup and more with just flatpak.
I think Linux sometimes has the same issue Microsoft does and that is if you tell people we want to move to new thing, it has better security or maybe just less awful for developers, and given the stuff X devs wrote about what drove them to well create Wayland or move to it I get that. The problem is if you just leave old thing there you get developers never putting in the effort to Migrate. Microsoft has wanted people to quit using NTLM, but even some tools from their own teams have hardcoded NTLM rather then use negotiate.
It has not been until GNOME and KDE have started talking about ending x11 session support(they still support it via xwayland) and Fedora and other distros disabling x11 session that it feels like the real gas has been hit on Wayland and supporting it.
Honestly I think had Fedora not moved to it by default we would have kept moving way slower.
I don't have the ideal answer here, keeping the old stuff around and compatibility mode often causes no one to migrate. If Microsoft has announced server 2019 would ditch NTLM, no more support for .net in it or Windows 11. I think suddenly companies would find the dev time to ditch those features. Turning things off by default signals to the developers who maybe don't keep up on those features or understand they are deprecated that something needs to change as it does not work out of the box.
Going slow often results in no one moving over, but moving fast pisses people off and often the new solution does not or can't replace every feature. Kerberos due to its security features can't enable certain insecure practices that NTLM did for example. We can't for example enroll users print cards at the printer directly anymore because the printers can't be domain joined and can't delegate creds. We had to move to handing out cards already attached to peoples accounts.
Now those situations i don't think are 1:1 with this situation, but a lot of these features maybe could have happened years ago had there been a bigger push.
Like I said i don't have the fix or the solution to the ideal rate of deprecating and removing vs supporting things for compatibility.
FattyDrake@reddit
Wayland is flexible and is organized so new protocols can be added.
The only reason they were "omitted" was likely because not enough devs were asking for them.
If KiCad tried to port their app over to Wayland a few years ago they could've raised these points and they could've been added and fixed a lot sooner. It's only a problem now because the issue is being forced by distros (finally). It's the only way to overcome the inertia to just say "use X11 instead." That's quickly becoming not an option.
SeeMonkeyDoMonkey@reddit
I think it's fair to say that unrestricted access to those features/data is ommited by design - as one of the advertised improvements over X11.
The current work to provide that access in a managed system should plug the functionality gap.
It's a shame that it's taking a long time, but sometimes that's what happens when coordinating multiple groups with no central authority to impose decisions unilaterally.
Drwankingstein@reddit
I really wish xwayland was better, It's close enough that I would just use it, and hope fractional scaling isn't too busted.
FriedHoen2@reddit (OP)
Xwayland cannot solve the problems listed in the post.
gmes78@reddit
Like the other person said, rootful XWayland would have none of these issues.
Drwankingstein@reddit
which ones?
FriedHoen2@reddit (OP)
Pretty none of them.
Drwankingstein@reddit
now thats just an outright lie
FriedHoen2@reddit (OP)
You didnt read the original post on kicad blog.
tonymurray@reddit
I don't understand... The problems are specifically about running natively in Wayland.
TheOneTrueTrench@reddit
Rootful xwayland might help, depending on what you need.
grimacefry@reddit
Wayland is mostly developed by kids who haven't even been alive half as long as X11. It's been a decade long crawl to rewrite everything, amd reinvent the wheel, and make fuck ups all over the place, just for the sake of it, just because X11 wasn't cool enough. Imagine if the past 10+ years had been spent just improving X11 and making changes to its architecture.
QuickSilver010@reddit
fragmentation is such a massive issue of this
Mister_Magister@reddit
one step forward, two steps back
MatchingTurret@reddit
We discussed this last week: KiCad and Wayland Support