Wayland and BunsenLabs

unklar · 2024-02-07 08:56:18

PackRat wrote:

Had some time today so I installed Sid (Siduction) with Xfce4 and then installed labwc. One bug is that labwc also requires xwayland which was not installed as a dependency.
Siduction uses sddm as the login manager and labwc does install a *.desktop file show it appears on that menu.

I installed siduction-noX and updated the system.
Then I installed openbox and obconf, followed by labwc.
xwayland was installed as a dependency.

This error did not occur here.

I do not use a login manager. I simply start the system after logging in from tty1 with 'labwc'. I will create a corresponding entry in ./profile.

brontosaurusrex · 2024-02-07 13:47:19

Seem interesting: https://github.com/YaLTeR/niri

johnraff · 2024-02-08 07:29:18

That video I posted above is quite a good overview of the current sitauation of X11 vs Wayland. I'm sure searching around will bring up many more.

My take on 2024 from reading here and there:

1) X11 is a monstrous mountain of spaghetti code with some built-in insecurities just from the way it works, and some just from the huge effort needed to maintain it. Big players are starting to phase it out, and as the code gets less attention the bugs and security cracks will increase. It won't be this year or next year but eventually X11 will die.

2) There is a whole lot of stuff that people use that doesn't work on Wayland yet, and some of it never will. Some apps will have to be dropped and replacements found. Some people will just have to give up on their favourite things...

3) Xwayland provides a way to carry on using some apps, but I doubt if there's a remote hope of using it with big things like Wine. (OTOH according to @CyberGizmo above, Wine are working on Wayland support.) Also apps will run a bit slower on xwayland because of the extra layer that comes in.

4) There's no rush at all to switch to Wayland (agree with @hhh there), but it's there in the future. No harm to start getting ready.

Wayland for regular users

A couple of small advantages:
1) Native Wayland apps should run a bit faster because the code is cleaner. That's more noticable with old hardware such as what a lot of BL is run on.
2) There's better support for multi-monitor setups.

But not much more than that.

Wayland for distro managers (like us)

Plenty of extra work, finding substitute apps, different configuration etc etc. No particular advantage that I'm aware of atm. Maybe down the road...

Wayland for programmers

Lots of fun I'm sure.

It's interesting to compare with systemd, which brought lots of benefits for regular users, and system managers too. Plenty of opposition from some heavy-duty Unix people though. Wayland arrives from the same RedHat origins as systemd but seems to be getting much less opposition as such.

Here's XFCE's Wayland roadmap: https://wiki.xfce.org/releng/wayland_roadmap

cog · 2024-02-08 08:03:57

@johnraff that’s a good overview and theoretical comparison. Your spot on on the compatibility layers.

The realistic hardware results can be mixed and even more confusing.

For example my 10th gen intel thinkpad flies on wayland and makes me hate X11s performance whereas earlier chipsets fly way past mine on X11. It’s kind of crazy stuff when you actually go to run it on hardware.

Last edited by cog (2024-02-08 08:07:07)

cog · 2024-02-08 08:08:30

Probably as a distro maintainer the conservative approach is the right one when considering these results.

hhh · 2024-02-09 19:37:25

To cog's point about hardware, I also have better rendering OOTB with my Intel integrated UHD graphics. I need to create an xorg.conf to mitigate screen tearing on X11, after that the performance seems the same. But without the conf, tearing central.

https://en.wikipedia.org/wiki/Wayland_( … land_and_X

johnraff · 2024-02-10 05:01:38

hhh wrote:

My take though, if I were to distill it a bit... let's wait a bit and see?

cog wrote:

Probably as a distro maintainer the conservative approach is the right one...

Looking at all the work that still needs to be done, it's hard to see X being phased out for some time yet.

So for BL this likely means our main releases should continue to be based on X for Carbon and Nitrogen at least. Our choice should be based on what's best for our users, which certainly looks like X right now.

But when the day arrives when X dies and a switch to Wayland is inevitable, personally I don't want to be given the task of sorting it all out all in a month. That's why I'm also pushing for a side-project where we can get some of the questions answered in advance, even if some software choices might need revisiting later. Something like XFCE's minimal base, adding the rest slowly, bit by bit, as time allows.

The Wayland session in Xfce 4.20 will cover a minimally necessary set of features, with the intention of gradually adding missing functionality in preparation for subsequent releases.

https://linuxiac.com/xfce-4-20-will-keep-x11-support/

Something installable (netinstall script?) would also let adventurous BL users play with it, find problems and suggest solutions.

A small room at the back of the Lab, occasional sounds of breaking glass...

A name on the door? How about BeakerLand?

cog · 2024-02-10 06:08:31

hhh wrote:

Many applications are not Wayland ready.

@hhh really states the linchpin here.

I’ve got a program I depend on at work that will work under XWayland on gnome’s mutter or plasma’s kwin but not wlroots compositors like labwc in BL’s theoretical case or sway/hyprland in my case.

The reality is a distro is the starting point to run the stuff end users need to run and if it can’t do that then the wheels just fall off.

cog · 2024-02-10 06:09:33

@johnraff cool name / idea

or1o9 · 2024-02-10 12:33:16

@johnraff, thank you for initiating this, in my opinion, needed discussion. I see after reading everything three times that at this point I have nothing new or interesting to add to it.

So I ´ll just join the choir of wait and see how things turn out for now, and will also in the meantime reconsider my options.

johnraff · 2024-02-11 01:22:52

cog wrote:

The reality is a distro is the starting point to run the stuff end users need to run and if it can’t do that then the wheels just fall off.

This.
If BL doesn't provide a solid base then it's lost its reason to exist.

PackRat · 2024-02-11 13:34:35

But when the day arrives when X dies and a switch to Wayland is inevitable, personally I don't want to be given the task of sorting it all out all in a month.

This. Good to be thinking about it and developing a roadmap now.

hhh · 2024-02-11 18:55:49

https://wiki.xfce.org/releng/wayland_roadmap
https://gitlab.xfce.org/groups/xfce/-/m … tab-issues
https://wiki.xfce.org/releng/4.20/roadmap

johnraff · 2024-03-12 01:49:20

LXQt are working on it too.
https://9to5linux.com/the-lxqt-desktop- … land-ready
https://github.com/orgs/lxqt/projects/4/views/2
And at least some of the devs seem to be looking favourably at labwc:
https://github.com/lxqt/lxqt/issues/10# … 1759317897

PackRat · 2024-03-13 18:28:05

The developers of mate desktop have quietly released mate 1.28 applications with Wayland support

https://wiki.mate-desktop.org/developer … and-meson/

Noticeably missing is marco, the mate X11 window manager.

Found a git page where wayfire is used as the compositor for mate-desktop running on Wayland. I suspect a similar setup with labwc as the compositor with mate tools would be doable.

So many experiments. So little time.

hhh · 2024-03-13 23:27:09

PackRat wrote:

So many experiments. So little time.

Fedora has announced dropping their xorg session...

https://news.itsfoss.com/fedora-41-to-drop-xorg/

https://www.theregister.com/2024/03/13/ … s_x_gnome/

https://pagure.io/fedora-workstation/is … ent-899128

But I don't think there's any danger of Debian doing the same for their GNOME or KDE sessions in the next release.

PackRat · 2024-03-13 23:58:45

hhh wrote:

PackRat wrote:
So many experiments. So little time.
Fedora has announced dropping their xorg session...
https://news.itsfoss.com/fedora-41-to-drop-xorg/
https://www.theregister.com/2024/03/13/ … s_x_gnome/
https://pagure.io/fedora-workstation/is … ent-899128
But I don't think there's any danger of Debian doing the same for their GNOME or KDE sessions in the next release.

That makes sense. Wayland is a RedHat project isn't it?

johnraff · 2024-03-14 01:19:14

Plymouth, pulseaudio, systemd and wayland were all started by Redhat, right?
Maybe GNOME too??

PackRat · 2024-03-14 09:59:55

Not Gnome. Gnome started as an independent project developed by a couple guys (both from Mexico, I think) because the Qt toolkit (KDE) at the time had a proprietary license. So they used GTK.

The rest are Red Hat.

Last edited by PackRat (2024-03-14 12:22:25)

brontosaurusrex · 2024-03-14 11:00:38

^ I thought GTK+ is is gimp related?
https://people.redhat.com/mclasen/Useni … s/x29.html

#21 2024-02-07 08:56:18

Re: Wayland and BunsenLabs

#22 2024-02-07 13:47:19

Re: Wayland and BunsenLabs

#23 2024-02-08 07:29:18

Re: Wayland and BunsenLabs

#24 2024-02-08 08:03:57

Re: Wayland and BunsenLabs

#25 2024-02-08 08:08:30

Re: Wayland and BunsenLabs

#26 2024-02-09 19:37:25

Re: Wayland and BunsenLabs

#27 2024-02-10 05:01:38

Re: Wayland and BunsenLabs

#28 2024-02-10 06:08:31

Re: Wayland and BunsenLabs

#29 2024-02-10 06:09:33

Re: Wayland and BunsenLabs

#30 2024-02-10 12:33:16

Re: Wayland and BunsenLabs

#31 2024-02-11 01:22:52

Re: Wayland and BunsenLabs

#32 2024-02-11 13:34:35

Re: Wayland and BunsenLabs

#33 2024-02-11 18:55:49

Re: Wayland and BunsenLabs

#34 2024-03-12 01:49:20

Re: Wayland and BunsenLabs

#35 2024-03-13 18:28:05

Re: Wayland and BunsenLabs

#36 2024-03-13 23:27:09

Re: Wayland and BunsenLabs

#37 2024-03-13 23:58:45

Re: Wayland and BunsenLabs

#38 2024-03-14 01:19:14

Re: Wayland and BunsenLabs

#39 2024-03-14 09:59:55

Re: Wayland and BunsenLabs

#40 2024-03-14 11:00:38

Re: Wayland and BunsenLabs

Board footer