Wayback 0.1 debuts as early Wayland server for X11 diehards Wayback 0.1 is out, the first preliminary release of the new Wayland display server whose announcement we reported a few weeks ago. "Just enough Wayland to make Xwayland work," reads the new site's handy strapline. It's a good summary – if you know what Xwayland is. This first, preliminary release arrives less than a month …
Worried about Wayland for myself anyway. Just recently "upgraded" (installed to another partition) to Mint 22 from Mint 20 a couple of weeks ago. My biggest issue was whether or not I would be able to get brightside(http://archive.ubuntu.com/ubuntu/pool/universe/b/brightside/) to build and run. Fortunately after not too much effort I was able to compile it and it's dependencies(about 17 of them, had to hack some things up to get some to build), brightside hasn't seen an updated since 2005(several other dependencies date back to early 2010s), and last officially released with Ubuntu 16 I think. Fortunately the app itself is super simple though it does depend on some more complex libraries.
brightside provides the MATE desktop (or Gnome v1 before that) the ability to flip virtual desktops when your mouse moves to the edge of the screen(up/down/left/right), something I have used as a core part of my workflow on Linux going back to the late 90s/early 00s with AfterStep. I use a total of 24 virtual desktops in a 6x4 grid.
Last time I checked, which was years ago at this point(when I first upgraded to Mint 20 from 17) I was unable to find newer software that provided this functionality for MATE. It seems the latest MATE includes Wayland support(Mint 22 ships with MATE 1.26 from 2021) , but worry/expecting brightside won't work with Wayland, maybe there is another software package now that offers this edge flipping functionality perhaps I am the last person in the world to use it or something.
My stopgap idea in the meantime once MATE 22 goes EOL(2029?) if there is no X11 anymore or something is get Ubuntu Pro, which looks like is free for personal use, which would give me another ~4-5 years of core updates anyway, since Mint bases most of their software off of Ubuntu's repos (I already maintain my browser updates completely manually with Firefox ESR and MS Edge).
Wayland has had very worried me for years because of this, I wasn't sure if my next jump from Mint 20 would have X11 anymore or not so was relieved that it was still there. By the time I'm done with Mint 22(which won't be till it's past EOL, excluding time spent with Ubuntu pro updates), it'll be a solid 30+ years of depending on edge flipping for my day to day workflow(and about 25 years of using a GNOME v1 style desktop UI), and obviously I don't want to try to force a drastic change in a workflow that has worked so well for so long for me.
From this point on, I will not upgrade any of my OSs until I have tested them on _all_ the applicaitons I use and all the software I have written (e.g. "legacy OpenGL).
The creeping sickness of systemd, wayland, gtk4/adwaita etc. will soon make "newer" Linux OSs a liabliity.
I am happy with Alpine 3.22 and MX-23.6
Just as a counterpoint, I was setting up a “kiosk” raspeberry pi this week, running a full screen browser. Did this the usual way: minimal install, then run xinit and a very minimal window manager (matchbox). It worked, but performance was very poor.
In desperation I tried following the guide from the raspberrypi.com site, which suggested starting with a full desktop install (which is wayland based), and to my surprise it was a massive improvement - fps at least tripled, at a guess - and quite a bit easier to set up too.
Just want to get that out there as a counterweight to all the grumbling.
This post has been deleted by its author
Yes, I know. I've done this before, but for whatever reason after installing xinit, xorg, all the usual gubbins to get it up and running, it worked but was clearly not accelerated. I needed to add or configure something, but damned if I knew what.
By contrast, trimming *down* the Wayland-based desktop was a doddle and left me with a clearly accelerated browser resulting in less CPU load.
It's not that it couldn't be done using X11, it's that it's fiendishly complex to get right. Wayland, in this instance, was trivially easy. Turning a full-fledged desktop into a kiosk involved editing /etc/xdg/labwc/autostart; one shell script. You can't say that about X11.
(As for the previous comments: Avahi is great - a few network packets every few seconds, and it makes managing hosts so much easier; and swap and /tmp don't come into the picture once you run overlayfs to ensure the SD-card is never written to)
This post has been deleted by its author
> brightside (http://archive.ubuntu.com/ubuntu/pool/universe/b/brightside/)
I am just curious: why the Ubuntu link instead of the source code?
https://github.com/dustincys/brightside
I am not sure if that's the original, mind you.
Looking, I also found this post -- was that you?
https://forums.linuxmint.com/viewtopic.php?t=328021
I found some other posts about it.
e.g.
https://ubuntu-mate.community/t/add-a-corner-activated-window-picker-for-marco/7814
https://ubuntu-mate.community/t/functional-funk-let-gnome-pie-enhance-your-gtk-desktop/22180
I doubt that Mint is going to help you much TBH, but if you can get the MATE developers interested in updating it, you may get a modernised version.
> Gnome v1 before that
Do you perhaps mean GNOME 2.x?
> expecting brightside won't work with Wayland
No, I'm sure it won't, but then neither will Metacity.
> once MATE 22 goes EOL
You mean Mint 22, right?
...this was purely a punne or play on words... [...] ...this is a Terry Pratchett reference...
Except that in the books I think it's spelled "pune".
Ah yes, here we are, first reference to it in the canon I believe, "Men At Arms", first page, Corporal Carrot's letter home to his parents:
"...we could inscribe on the back something like ‘A Watch from, your Old Freinds in the Watch’, this is a pune or Play on Words..."
Sorry to be the resident Ponder Stibbons.
You're lucky it's only a first offence, or you'd be strung up by your figgin.
> Begun the Linux war has.
"Use the force, Harry" as Gandalf said.
Linux is one big cluster of battles in a war that's now over 50 years old.
The trouble is that Linux folks are so obsessed that they haven't noticed that there are other armies battling alongside them -- the BSDs and so on.
They never realised that the leaders who started the war moved on to new battlegrounds. (Plan 9, Inferno, etc.)
It is beyond their ken that there are other territories that aren't even in their war: Lisp OSes, Oberon/A2/Zonnon, Tripos/AmigaOS/Helios, even their nearest neighbours on the islands of C++ (Symbian, BeOS -- even Haiku, although for short-sighted reasons they're building a bridge to the land that's at war)...
Come on Liam, I realize that you are joking, but claiming that people who have been reading tech news since the 1990s somehow "haven't noticed" is _some_ journalistic 'licence'!
It is pretty obvious that none of the alternative "battles" will be won, and Linux is falling to the legions of the clueless. An OS version of "eternal september". The trolls/astroturfers/newbies are already demanding systemd by default in MX. So much for choice.
They're selling hippie wigs in Woolworth's, man.
> claiming that people who have been reading tech news since the 1990s somehow "haven't noticed"
No. I stand by it, but I think you misread me.
Linux exists because GNU fscked up. The GNU project investigated and evaluated using the BSD kernel in the GNU OS in about 1988-1989 and decided against it.
If it had gone for it, there would have been a working FOSS Unix *before Windows 3.0* and the world would look significantly different.
The BSD folks screwed up by not spotting in '87 or so that the 80386DX PC was the future. It was clear enough to everyone else, but they've always fetishized arcane old stuff.
Unix is a religion, and like all religions, they're more interested in attacking each other's vile heresies than lifting their heads, looking at the wider world, and learning about it and how it works.
Even Thompson and Ritchie moved on. The acolytes still cling to the first revelation.
The proof is that so many of them still use Vi.
Whilst I agree that 68K UNIX was around in the '80s, I don't really think it was x86 Windows that killed it.
The main proponents of 68K UNIX moved on to RISC platforms, or left the market completely because they were failing to maintain their niche.
Outside of the world of Sun, SGI and Apollo, most 68K UNIX(ish, Domain OS was not actually UNIX) implementations were more about the multi-user aspect of computing than the GUI aspect. If you look at pretty much any of the representative systems running Unisoft UNIX that were generally sold into what is now small business territory, and were used concurrently by multiple people using serial terminals, these were in the in the same market as RSTS/E on PDP11s, low-end VAXes or the various IBM OSs on System/34 and System/36. This was never about GUIs.
What killed this market was actually not Windows, but the MS/DOS PC revolution. We had a stupid situation where an IBM PC clone with graphics, local storage (even if it were just floppy) and local software was not a whole lot more expensive than just a single serial terminal (and taking into account the shared cost of the department server itself made this even more the case), so many, many businesses shifted from small departmental supermicro/mini computers to disconnected PCs running Sneakernet (floppy disk based data transfer). IMHO, this was a huge backward step in the evolution of small computers. And on top of this was eventually bolted very simple and grossly insecure networks to aim to provide similar functionality to a shared system.
By the time Windows 3 came along, the 68K based market was already essentially dead, not necessarily because of performance issues but because of what companies were buying, and the largest players who weren't into the departmental super-micro had moved to UNIX on RISC, because UNIX and RISC fitted together very well, with high power, high cost UNIX workstations, and systems that scaled all the way up to mainframes.
What Windows did was hollow out the UNIX workstation market in the '90s, really leaving UNIX in the medium-to-large system market, and Linux occupying the capable PC niche, firstly for enthusiasts, then for network servers.
I don't doubt that if Motorola had the 68000 and it's associated chipsets ready at the time that IBM was looking to design the IBM PC (and that they could have got IBM's attention - remember the IBM PC was an under-the-covers stealth project that was not really sanctioned by IBM management), we would have been in a much better place much earlier, and even though the first 68K family implementations were missing some memory management and protection features, so would not have run UNIX very well, we could well have seen UNIX (and maybe some other multi-user, multi-tasking OSs like Concurrent CP/M)on the desktop sooner and more often than we actually did, although we would probably have to have seen a large change in the UNIX licensing model to make that more affordable.
You could run Xwayland rootfull and fullscreen already on the existing compositors. I did this in crostini with ChromeOS's compositor, but it should work on sway, gnome and plasma. Then I ran my x11 windowmanger of choice, evilwm.
I'm pleased wayback is getting momentum but it's only useful if you want ultimate lightweight layers/processes running (if Wayland+typical DE is bloated and slow on anything built in the last 15 years I'll be surprised and disappointed), or if at some point gfx drivers become wayland only but you want to remain very old school and/or still had an unacceptably weak machine. I'd assume just using Xorg is going to be better for years.
Wayback implements the bare minimum to get Xwayland functioning, so if you wanted to run a wayland only app like waydroid you'd not be able to just flip to the main Wayland compositor for that.
Now if it had future plans to allow your x11 windowmanager to move/resize rootless Wayland apps that would be amazing, but I don't see any mention.
> You could run Xwayland rootfull and fullscreen already on the existing compositors
I'm aware.
Original XFCE creator Olivier Fourdan is now at Red Hat and has blogged about it, as I mentioned here:
https://www.theregister.com/2023/10/27/mint_wayland_cinnamon6/
I mentioned his efforts in my previous story on Wayback:
https://www.theregister.com/2025/07/03/wayback_wayland_display_server/
(I misspelled his name. _Mea culpa._)
The blog only got 3 posts, alas. It is here...
https://ofourdan.blogspot.com/
It was explicitly experimental.
AIUI the aim of Wayback is to make this as simple, and requiring as little config, as possible. That sounds laudable.
I think supporting Wayland-only apps in X11 desktops falls outside the scope of what Wayback is looking to provide. Giving distros a pathway to be able to drop the Xorg server without orphaning some treasured X11-only DEs is the name of the game; throwing those DEs a further lifeline of support for apps that have left X11-native behind is manifestly not.
To their credit*, the XLibre mob do seem to be motivated to address this, with discussion of integrating an embryonic tool called 12to11, which appears to be akin to a "reverse Xwayland":
https://git.linuxping.win/12to11/12to11
[*Albeit that the stated motives range from the practical utilitarian to a desire to "own" any app devs who've had the temerity to abandon native X11 support. That place is giving my popcorn dealer aneurysms.]
" something really legacy like the Common Desktop Environment, CDE."
Tsk! Tsk! The only one true window manager is ol(v)wm - OpenLook (Virtual) Window Manager (See: https://en.wikipedia.org/wiki/Olwm)
The later window managers (including CDE) are just Splitters!
Now, where's my slippers and pipe...
