Yserver: A modern X11 server written in Rust
github.com
github.com
by Mark Thomas as an experimental sucessor of the "X Window System" (its development has been cancelled for a long time; the latest release that is available on this website is from 2004).
The German Wikipedia still mentions the "Y Window System":
> https://de.wikipedia.org/w/index.php?title=X_Window_System&o...
At that time, what should replace X11 was a heavily discussed topic, so people who were already interested in programming under GNU/Linux at this time typically remember intense discussions on the internet about the pro-s and con-s of lots of proposals to replace X11.
Proposals that I have in mind are:
- Y Window System
- Fresco: https://web.archive.org/web/20100729184325/http://fresco.org...
- Xynth: https://github.com/alperakcan/xynth
- Xfast: https://sourceforge.net/projects/xfast/ (here a German article about Xfast from 2008: https://www.linux-magazin.de/ausgaben/2008/11/ohne-x-tras/ )
- Mir (Ubuntu) as the main competitor to Wayland: https://en.wikipedia.org/wiki/Mir_(software)
- Wayland
- ... and very likely many, many more proposals.
See also https://en.wikipedia.org/w/index.php?title=Windowing_system&...
No one's comparing this to 'y'combinator or 'y'ahoo.
'server' and 'window system' are completely different...
I'm curious why multiple screens is considered legacy baggage and thus out of scope, given how common multiple monitor setups are these days. I also have zero familiarity with X internals, so don't know if multiple monitor support is a horror show that'd be miserable to support.
Xinerama (the extension that enables one virtual screen over multiple outputs) existed before but the layout could only be defined statically - so you'd need to restart your X server with a different config if you wanted to connect a monitor or a projector or something.
I can't recall any application able to use multiple X11 screens (except in the trivial sense that you could set DISPLAY when starting the application), and I've been using X11 since X11R4.
[1] https://en.wikipedia.org/wiki/Xinerama [2] https://xorg.freedesktop.org/archive/X11R7.5/doc/man/man1/xr...
Traditionally each screen in an X11 setup was it's own separate thing with it's own separate frame buffer. While technically applications could move between screens, this depended on the application caring enough to do so. It had to maintain two(or more) mirrored windows(one per screen) and keep them all aligned. So realistically no application did this.
The modern method of doing multi monitors on X11 involves one large virtual screen with each monitor assigned a section of it. This has downsides, for example; this is where the myth that X11 can't do mixed DPI setups comes from. But it has one huge massive overwhelming upside. The application does not have to be aware that there are multiple screens and multi monitor setups just work.
What most X11 users actually do is set the global DPI to that of the highest DPI monitor, and use xrandr to scale down the framebuffer of the lower DPI monitor, which "zooms it out". Note that this has performance and image quality implications. There's a guide on how to do this here: https://blog.summercat.com/configuring-mixed-dpi-monitors-wi...
Admittedly X11 mixed DPI was using separate screens which were awkward to deal with and early versions of the unified screen tech (xinarama and xrandr) did not support mixed DPI. And even modern X11 while it provides the needed DPI information requires the application to care enough to support it. Which really means unless the toolkit provides it for free most applications are not going to do anything,
I'm not sure what you're talking about, fractional scaling is just another way to describe DPI. The scale factor is just the DPI divided by 96. If you're referring to windows getting scaled by the compositor for fractional scales, that's only used for older software. Both Qt 6 and GTK 4 support natively rendering window contents at fractional scales on Wayland.
Wayland only supports fractional scaling. and there is a good argument that this is a better system, not because it looks better, it does not. but because applications don't have to be aware of it where a mixed dpi requires the application to actively deal with it.
I've tested per-monitor DPI before, and [0] mentions one way to do it. I tested per-monitor refresh just now. Using the xrandr CLI to set the refresh rate to 24.0 on my primary monitor and 60.0 on my secondary results in "cinematic" visuals on the primary monitor and "soap opera" visuals on the secondary.
I'm currently using Windowmaker, but I see no reason why this wouldn't work with KDE.
It has been nearly eighteen years since the Wayland project started, and they are still not at feature parity with the major windowing systems. [0] It's nuts how long it's taking them. [1]
[0] As one example, apparently the Wayland policy for clients that stop responding for a few seconds and fill up their event mailbox is still to terminate the stuck client. If memory serves, Windows 9x handled stuck clients better than that.
[1] I'm sure it's good enough for what you're doing and you never run into any rough edges or misfeatures, so don't bother chipping in with that retort. ;)
I haven't had complaints there for Wayland but I will say that it breaking other things has been annoying.
Using the xrandr CLI to set the refresh rate to 24.0 on my primary monitor and 60.0 on my secondary results in "cinematic" visuals on the primary monitor and normal "soap opera" visuals on the secondary. Setting the refresh rate back to 60 on my primary results in "soap opera" visuals on both.
I'm currently using Windowmaker, but I see no reason why this wouldn't work with KDE. I'm using xorg-server 21.1.23 (which supports RandR 1.6), xf86-video-amdgpu 25.0.0, xrandr CLI version 1.5.4, and kernel 7.0.12.
I'm on Gentoo Linux. I would not be surprised to learn that Debian (and Debian-derived distros) never shipped a version of Xorg or the related libraries where this worked correctly.
I'm personally very dubious of the claim. This has basically never been supported because x treats all screens as one big screen unless you run multiple X screens which disallows moving Windows between screens which is a pretty big barrier to normal usage
amdgpu + x11 + xfwm4
$ xrandr | grep -A1 ' connected' | sed 's/^/ /'
eDP connected 1920x1200+0+240 (normal left inverted right x axis y axis) 286mm x 178mm
1920x1200 60.03*+ 40.02
--
DisplayPort-0 connected primary 2560x1440+1920+0 (normal left inverted right x axis y axis) 597mm x 336mm
2560x1440 59.95 + 200.00* 179.96 144.01 120.00
Is it? Or is it forgotten?
Around 1994 I had a Pentium 133 with 16MB. In it a Diamond Speedstar 24(Pro?) (Tseng ET 4k) Vesa Local Bus, some ISA Trident 8900, and an ISA Hercules, driving one 17", one 15", and the Hercules at something like 12" IIRC. One could choose in the BIOS which "GPU" ...err... frame buffer should have priority at boot, or rather which slot, so when you've chosen VLB it took that, and the others were a matter of the OS to initalize and drive way after boot.
At the time I compared 386BSD, NetBSD, SLS(Softlanding Systems), early Slackware and SuSE, and lo and behold, I could move windows across all of them on all of them!1!!
With proudly created custom modelines for all of them, even the Hercules, with different Hz & DPI for each screen.
Though it didn't really make sense, because X on the Hercules was very laggy and jerky, coz' 8-Bit ISA. Was more useful for syslogs and debugger.
Anyways, it worked, even if only as POC, to show off.
Now that wasn't Xorg, but XFree86, but still?
1994. Worky, worky!
IIRC that also applied to Accelerated-X, at least for the Tseng and Trident.
Didn't try the Hercules with X then.
Yep. I remember that pride very clearly. I'm also so glad I never have to do that math ever again.
> SLS(Softlanding Systems)
I'd never heard of this one. If the claim that its slogan was "Gentle Touchdowns for DOS Bailouts" is true, then that's a really great pairing of distro name and slogan.
Isn't that fun? https://nyanpasu64.gitlab.io/blog/crt-modeline-cvt-interlaci... :-)
TBH I've been mostly unimpressed with the early Linuxen. At the time they had almost nothing which NetBSD didn't have, and I knew that really well. Gentoo 'ricing' well.
But it was clear that there was momentum behind the Linux hype. It was on CD-Roms in magazines, in book stores. While *BSD wasn't known generally, just by some guys in Universities, or similar. Bad marketing. FUD because lawsuits, and whatnot else.
Such BSD, so sad...
May I submit that it is more likely that you are speaking of 1998 instead of 94 rather than this entire technology working differently.
No. I'm not pulling this out of thin air, or misremembering. It wasn't easy, and it wasn't in the manpages, or only some of it. It was rather involved, and didn't work on first try. Not only that X-thing, but also which TTY/VT was on which head for tailing syslogs, std-error, and whatnot else. But it worked. And with no X on the Hercules, just between the Trident and Tseng, even reasonably fast.
I remember exactly because I moved shortly after that. I also remember which relief Xinerama brought me, when it appeared :-)
Edit: I also remember frying the Hercules and the attached screen with both of them giving me the magic blue smoke, because I've overdriven them a little. But it didn't matter, because that was already abandoned cybertrash at the time, used, and collected just for fun :-)
Hmm.
I'm not sure what windowing system you've been using over the past ~thirty years, but you seem to be unaware of both Xinerama (released in the late 1990s) and XRandR (version 1.2 [0] of which was released in like 2006). Maybe you've been using an X11 implementation provided by some proprietary *NIX for all of these years, but whatever you've been using, it has certainly been neither XFree86 or xorg.
> The only exception is that monitors that support variable refresh rates may be able to offer this feature in multiple monitor configuration subject to software and hardware options.
1) Monitors have been able to support multiple refresh rates for ages. This is a big part of why EDID and friends exists.
2) I note that VESA standardized Adaptive Sync in ~2009, and that VRR-supporting monitors were extremely uncommon in the consumer space until the introduction of GSync and the addition of Adaptive Sync to the DisplayPort standard... which both happened in ~2014. Add to that the fact that my secondary monitor does not support VRR, and it becomes very clear that VRR is not a prereq for driving multiple monitors at different refresh rates.
[0] IIRC, 1.2 is the version that gave it feature parity with Xinerama
Old versions of GIMP (back when the toolbars etc. were separate windows) used to let you move any of its windows to a different X screen. And by "move" I don't mean drag - there was a menu where you could select the screen to move to.
I have to wonder if the fact that Wayland either never had or has only very recently gotten support for applications that need to place their windows at application-commanded locations on the screen meant that those lovely tear-off menus had to die.
Menus are not meant to be torn around.
Yeah, they are meant to be implemented with web technologies and look like shit.
BTW, this tear-off style is probably quite old. Long ago, I used an early version of ANSYS (for Windows) which apparently was still close to its Unix original, and it had its menus pop up like real windows, with close buttons! They were nicely cascaded, but one could rearrange them.
Yeah, I agree with that. I was using some ancient X11 program that had tear-off menus, but I'll be fucked if I can remember which one it is.
> Yeah, they are meant to be implemented with web technologies and look like shit.
Yuuuuuup. If you always take the "yes" side, you'll come out quite a bit ahead of your fellow gamblers for the "Will GNOME make things worse for sophisticated users and call it 'simplicity'?" wager.
And that’s understandable. It’s not actually good usability.
They're really great on systems that let you hold a modifier key and then a mouse button to drag windows around... rather than requiring you to find the very-small-compared-to-the-size-of-the-entire-window portion of the window you can click to change the window position. They get even better when you're on a system that reliably remembers the position of application windows.
Folks who have never used a system that lets you relocate and resize windows without first moving your mouse cursor to "blessed" regions of the window absolutely do not know what they're missing.
> ...they have been unfashionable...
Fashion is for people who love doing busywork. Where fashion gets nasty is when that busywork makes a bunch of work for everyone else.
There is also wors.
I genuinely wonder if they stopped to think why X11 has sockets or just blindly copied it over. Or are they unaware other forms of IPC exist, that don't require you to go through the kernel 13 times to send a byte to the other process?
UNIX sockets are perfectly fine for IPC with small amounts of data, and is how everything in UNIX has always done it, network transparency or not. They provide a simple, efficient and reliable communication channel between two processes.
A long time ago when I looked at designing a X11 replacement, that was my approach. AFAIK, only special X utilities used anything but Xlib anyway. And later I think this is what early revisions of Canonical's Mir did.
So what would you do differently in an alternative client library?
I should have better disclaimed my comment.... to be clear I don't know much about the graphics subject, I probably should have prefaced it with,
"I don't know anything about Wayland but as someone totally naieve on the subject but assuming someone else's assumption".
At least to me, even if it breaks the X11 model (Which is a shame, that was fun to play with back in the day) if they're doing it the way they are I'm guessing Chesterton's fence will come into play at one point or another.
1. You don't need to convert your discrete messages into a stream with size metadata, only for them to immediately be converted to a message on the other side.
2. You don't need to jump into the kernel to copy over 20 bytes, only for the other side to jump into the kernel to copy it back.
3. You don't need to deal with the "oh but what if my read returns half a message because this is a stream"
4. You don't need to pretend you're doing network programming.
It's pretty representative of the project though - "We're doing things the way we've always done them, but slightly different. Now rewrite all your software to work with our thing. No, you cannot do global keyboard shortcuts or set window position. You don't like it? We're doing this for free, you cannot critique it."
You need some kind of serialisation either way. It can be as simple as “this message has the shape of this C struct”, but that’s the case whether you’re talking shared memory command buffers or sending data over a socket (and there are good arguments for and against in both cases).
You’re right that you don’t need to deal with “oh I received half a message” when using shared memory command buffers, but that’s more a code complexity thing someone solves once in wayland-client and then nobody has to really think about it again. It’s not really a performance concern (because hopefully the rx buffer is large enough for it to happen rarely) or application code complexity concern.
If the compositor wants to notify the client as soon as possible, it has to send 1000 messages per second. If you buffer them, you're wasting the hardware's potential, if you don't buffer, them you're doing 1000 write()s per second, which is... ugh.
If you're literally going to design the protocol from scratch and require all existing software to deal with it, why not pick the IPC model that doesn't have this issue.
Or would the receiver spin in a tight loop on a memory load from some byte in shared memory which indicates a new buffer is submitted, so that it gets notified without involving the kernel? Or is there some fancy mechanism I’m not aware of?
I would probably expose a poll() and let the client deal with it, I don't know if there's a one-size-fits-all signaling mechanism. But you have control over it, which is probably another plus.
Speaking of... it looks like common Wayland compositors [0] still kill clients that can't keep up with "high speed" event generators like 1kHz mice. [1] So, that's nice.
(For people who plan to retort with "just handle events in a timely manner", check out the comment here [2]. OSX, Windows, and X11 all cope just fine with programs that go unresponsive for multiple seconds. If the statements in this bug report (and the reports I've read elsewhere) are accurate, Wayland doesn't... and that is inexcusable.)
[0] ...or whatever the Wayland terminology is for the thing that does the work of the X11 compositor + window manager...
[1] <https://gitlab.freedesktop.org/wayland/wayland/-/work_items/...>
[2] <https://gitlab.freedesktop.org/wayland/wayland/-/work_items/...>
Sounds like a security nightmare.
The kde one doesn't support remote user login, and while the gnome one does on paper, I never got it to work. The remote connection situation in Wayland is a major regression.
waypipe ssh XXXThe people who put together TS/RDP are geniuses IMO, it's insane as to how usable it has been for at least 15-ish years...
Security-wise there are concerns but...
Early dial-up Internet days (early for me), 28.8k modem, I was already running Linux, probably on a 486. I also had a very old PC laptop (I think a friend of my parents gifted it to me after he got a new one from work), a 386 I think (with the horrible slow display/refresh rate: a TFT IIRC). I used a parallel cable and PLIP (Parallel Line IP) and X11 networking to send a window manager+browser from the desktop (the 486) to the laptop.
So my brother and I could both go on the Internet at the same time.
It felt like the future and, honestly, we've kinda seriously regressed when it comes to "GUI over the network".
These days it’s unbelievably niche, as opposed to more controlled screen sharing scenarios. I think it’s understandable that it doesn’t get a high priority.
Looked nice, but crossed it off as soon as I saw that, as I'm working on a project currently that uses many screens. Can't just call a thing legacy because you and the people you directly know aren't using it.
Nowadays, multiple monitors present one big virtual framebuffer and only one logical X11 screen.
The normal, usable way to have multiple monitors for your X11 desktop environment is for them to all be combined into one logical screen that you can move windows around in, and that applications aware of the right extensions can discover the actual physical layout of the monitors that comprise the single logical screen. Multiple screens in that X11 sense is a far more obscure feature than simply supporting more than one physical monitor.
What they are talking about is supporting more than one of those, and from app's perspective they are completely separate (can't move windows between them).
While I can see the use cases (say secondary screen only running single app) I never actually used that feature so it's understandable drop.
edit: AND i've been using GNU/Linux and derivatives for the last 20 years.
At least it is worth reading.
- [0] https://wayland.fyi/
Something like XLibre or Phoenix would have been taken very seriously 5 years ago.
Xorg worked under nvidias for years.
And for example, it's a weird signal to me when somebody believes the reason X11 has baggage is because it does byte swapping for endianness. This statement alone taints the entire rationale for the project.
There is no need to put this code on GitHub. Everyone with an API key can achieve the same if you hand them the prompt.
This is like committing build artifacts to version control.
On top it's such a lame idea. "What if rewrite in rust applied to X server". Fits on a napkin. Man what a nothingburger :(
Would that change anything about the fundamental cliche-ness here?
Also, no, I'm not clever, but not sure what that has to do with this comment chain.
Look, I've been writing open source software for 20+ years, and after getting seriously burned out by it, I picked it up again with Claude (proof: https://github.com/jleclanche)
I can tell you a few things from that:
1. I'm writing better software than before, because AI is less lazy than I am. It's not necessarily always smarter, but writing correct software has gotten so stupidly cheap that it doesn't make sense not to do things right... so when you tell AI to do things correctly, it tends to know what you're talking about.
2. I'm more curious than before, because AI gives me time to explore many paths, very fast. A project like this one, like someone else said elsewhere in the comments, is more about the journey than the destination.
There is no "write me an X11 server but do it in rust and post on hn" prompt that does the thing. There's a journey of building, learning, understanding.
I'm not saying the resulting software is particularly valuable, but the journey is. This is HN, and you're shitting on someone who is using the most powerful pieces of technology we've achieved to go on a journey of discovery of X11 internals for the past 2 months. It's just shameful.
And yeah, if I were the author, I'd run claude over all the transcripts and extract a story with what's been taught and learned throughout. But I'm not the author. Just someone enjoying living in absolute science fiction.
i'm personally okay receiving presents on my birthday even if they were purchased from a store on the internet, and i'm okay receiving software presents on github.com even if they were purchased from a store on the internet.
Also this is slop.
Is anything done with AI automatically slop? I don't understand this