Canonical Reportedly Not Planning To Enable Wayland-By-Default For Ubuntu 20.04 LTS

[Creak]

Hm, both of these work for me. What exactly doesn't work in Nautilus?

I definitely had the problem several times, but I'm not sure of the exact repro steps. Try opening two nautilus windows and drag and drop a file from one to the other.
(I'm on Fedora 29)

[Chugworth]

GNOME's Wayland support is one thing I'll give them credit for. I was testing it just a few days ago on Arch, and it felt pretty solid from what I could tell. But as well-polished as GNOME feels, the interface just sucks in my opinion. I couldn't take more than just a few days of it before I switched back to KDE Plasma. I'll admit though, KDE's Wayland support is pretty awful right now.

[R41N3R]

GNOME's Wayland support is one thing I'll give them credit for. I was testing it just a few days ago on Arch, and it felt pretty solid from what I could tell. But as well-polished as GNOME feels, the interface just sucks in my opinion. I couldn't take more than just a few days of it before I switched back to KDE Plasma. I'll admit though, KDE's Wayland support is pretty awful right now.

If you know how to work around the Wayland issues of Plasma, it works very solid. I use it every day. But yes, there are some serious issues that should have been addressed a long time ago like the rendering glitches. But why do you think it is awful? I would say it definitely needs much more attentions by the developers but it is not that bad.

[calc]

So the primary two reasons for Wayland to exist are(?):

1. Being able to run random malware downloaded off the internet 'securely' instead of just using verified good software via your distribution.

2. Being able to support Gnome's version of fractional DPI scaling, KDE's version works just fine with X11.

Bonus feature:

3. Any time your window manager dies it takes your entire desktop with it, unlike what normally happens with X11. And you can't restart your window manager without logging out, eg Gnome extensions can't be reloaded.


I definitely can see why everyone wants to move to Wayland en masse.

Wayland is a 'solution' in search of a 'problem'.

[bwat47]

People keep saying that that's a G-S on Wayland issue, but I just had a Plasma on X11 crash and my entire session was gone, had to re-login. So it happens on X11 too (maybe not with GNOME Shell, but at least with Plasma, as I just proved).

I don't use KDE so I can't comment on how it behaves.

Gnome on xorg recovers cleanly from gnome-shell crashes (the shell just disappears for a second and then re-appears. the same behavior as if you restarted gnome-shell by going to alt + f2 and using the r command)

If xorg itself were to crash, it would take the entire session (regardless of DE), but I've never experienced a xorg crash in the entire time I've used linux.

I've definitely experienced gnome-shell crashes (and gnome-shell on wayland encompasses so much; the display server, compositor, window manager, shell all in one process. It would be pretty hard to make that 100% crash free, and a crash in any of this functionality will take down the entire session, so I'm pretty concerned about the lack of crash recovery with this design).

[hax0r]

I still remember 10 years ago when Canonical said 12.04 would ship with Wayland by default.

[dungeon]

I still remember 10 years ago when Canonical said 12.04 would ship with Wayland by default.

Right, but then GTK3 apocalypse happened and the rest is the history

[vegabook]

That's pretty funny considering Mir actually failed entirely, while Wayland is still being actively used and planned to go out.

Wayland is the future, and always will be.

It's failing because it doesn't offer enough that X doesn't already offer. The only thing actual end users care about that it offers, is hidpi scaling, and that's not enough to make people want to put up with its bugs. It needs to WOW people so that they'll adopt even with its problems, thereby creating a virtuous loop. For now it's a perpetually-delayed, spluttering refactor that's at any time in real danger of being left behind by something revolutionary based on Vulkan or something that might actually leave the whole windowing paradigm behind.

Even its signature pull, per-monitor scaling, might vanish once all monitors are hidpi.

For 1990s era UIs, namely windowed desktops, X is actually fine. At least Mir tried to do something big with the phone. That it failed does not by default make Wayland better, or indeed necessary.

The first rule of software success is that you've got to create user pull. Pushing a buggy and feature-deficient refactor, doesn't work no matter how elegant it is underneath.

It's OS/2 all over again.

[dungeon]

Wayland is the future, and always will be.

Bullshit, someone will write Zeelönd as replacement for old Wayland sooooner or läter

[Creak]

Bullshit, someone will write Zeelönd as replacement for old Wayland sooooner or läter

Of course there will be a replacement for Wayland one day, but considering the time it took to get Wayland to where it is now, I don't think that if a new project were to arrive next month with a goal to replace Wayland, it would have a lot of pull from the devs. The paradigm shift from X.org to Wayland is huge and that is why it is taking so long. It's not just a rewrite of an old software, Wayland took some design decisions based on more modern window systems which needed to offload a bunch of code on the DE. That is basically what took the community so long to get there.

To me, Wayland is definitely the future we're going to have at least for the next decade. After that, I don't know, but the refactor that would have been done by the DE in order to run on Wayland might help them to port on whatever X server that could happen later (like one with a mandatory dependency on Vulkan, for instance).