Xfce's Wayland Compositor Code Continues Improving

[muncrief]

Since Gnome and KDE went down the Windows 8/10/11 route to ludicrously over complicated and inefficient GUIs I've found XFCE to be the only remaining fully functional DE dedicated to getting work done with the least mouse clicks and scrolls, and keypresses.

I mean seriously, remember when GUI efficiency used to be measured with these simple metrics? And the majority of Linux users scoffed at the inefficient, kindergarten like, Windows GUIs?

I do.

And then within a few years most Linux DE developers suddenly began emulating the ridiculous Windows desktops. Sometimes with even more kindergarten like spinney shiny things.

It still boggles my mind.

But hey, the beauty of Linux is that users can choose whatever DE they want. Even if others disagree with their choices.

[asriel]

Hey, I'm just some guy that held out until sway could provide me with a fully functioning desktop. I don't see how that's egocentric or a megalomaniac.
If you have other needs and they are not being met on wayland by whatever you're using, that really isn't any reason to go around labeling everyone else.

Delete Xwayland from your computer to run PURE wayland and you will soon understand that wayland is crap. Still there is no other way - all money are invested in wayland and sooner or later it will become the only alternative. So yes - it is good that Xfce is trying to support it and make our shitty future stink less. The only point I do not understand - why one more compositor??? A lot of projects doing their compositors and all of them are kind of unusable so far. Only Gnome is more or less functional, and if you hate Gnome at least half a way I hate Gnome - then there is a very hard choice. Nothing works. Something works here - something works there. The most logical idea would be to concentrate on wlroots based development and start contributing to wlroots so it will become something that can work, but instead of that there is a zoo of different compositors. Useless, buggy and ugly.

[Monsterovich]

Nice of XFCE joining the 21st century.

Why does your 21st century look like hell?

[Monsterovich]

Why don't they use wlroots though?
It's quite mature compared to weston and can create a very much usable desktop (swaywm).

This is a question to ask GNOME and KDE developers. In fact, the Wayland architecture is broken from the beginning and wlroots doesn't fix this.

[mrg666]

The transition is picking up on the desktop from X to Wayland. After LXQT, good to see XFCE is enabling Wayland as well. Plasma and Wayland has been working just fine for me since more than a year now using an AMD GPU.

[fahrenheit]

As far is public the plan is to use wlroots, see https://wiki.xfce.org/releng/wayland_roadmap (also check the page history for discussion, reasons and discarded options)

[Steffo]

Wayland is reference protocol, Weston is the "proof of concept" implementation. Weston is not meant to be used as a daily gear, but as an example implementation of Weston protocol features. Or am I missing something?

Wayland is a reference implementation. There is no tag saying: "DO NOT USE ME". What sense would this make?!

[eagleoneraptor]

Delete Xwayland from your computer to run PURE wayland and you will soon understand that wayland is crap.

Why deleting XWayland? It was part of the Wayland ecosystem from almost the beginning. Why being so dogmatic?

[arQon]

The only point I do not understand - why one more compositor??? A lot of projects doing their compositors and all of them are kind of unusable so far.

Because that's the whole point.

The original promise was that Wayland would provide a complete working system, the way X does, such that the DEs only had to build the layer above that. Then someone at IBM realized that if they didn't deliver on that after all, and instead forced every DE to write half of the "Wayland" code themselves, it would more than double the workload for everyone but IBM.
The sensible response would have been for the remaining DEs to co-operate on a shared compositing layer, but timing and manpower issues made that an unlikely proposition, and as you can see from the current shitshow it's one that didn't happen.

Just because the DEs are (generally) free as in beer doesn't mean the commercial ones don't care enormously about market share. Causing as much collateral damage as possible is absolutely the end result that IBM was hoping for, and it remains a competitive advantage for them regardless of the numerous unintended Wayland failings that are still keeping *anyone* from being able to deliver a properly functioning desktop with it rather than just "not IBM".

The end is sort-of in sight though. Based on the last 15 years of "progress", Wayland should be "ready" in about another 5 years, at which point the CADT will immediately declare it "obsolete, insecure, unsupported, and unfixable", and start work on its replacement.

[arQon]

Why deleting XWayland? It was part of the Wayland ecosystem from almost the beginning. Why being so dogmatic?

His point is that XWayland is the *only* thing making "Wayland" desktops actually usable - which obviously makes sense, since nearly everything is still written for X.