NetBSD On The State & Future Of X.Org/X11

[User29]

E.g., I make use of shading windows (rolling them up into just a titlebar, with said titlebar left floating around the desktop)

OK, this is a good one, although I never used such eyecandy (but not stating that's useless because I don't use it). However I was pretty convinced that this "decoration madness" (rotating cubes, etc.) died out counless years ago.

TBH I'm a bit sad that openbox won't work on wayland (I have my "standardized" .config/openbox dir carried around from machine to machine in the last ~15 years) but I think that's the fault of openbox devs and not the X11 > Wayland switch and I've moved on.

[User29]

It doesn't matter what we want, or what the technical aspects are, practically speaking, it's the future.

Even better, it's the present. Those who left out will be left behind.

If I want to go sarcastic (I don't), I'd say this is the once in a lifetime (and probably the very last) opportunity of BSDs to get a visible amount of new users, offering the good old solutions for those who wants to stuck in the past.

There are two major problems with this approach, however:
1) even BSDs have started implementing wayland (what a treachery!!!)
2) serving the need of these new users would stop the development (advancement) of BSDs gluing them to the past which is definitely not what an OS developer wants to hear.

So yes, the future is here to stay.

[mrg666]

At this point in the face of inevitability if you haven’t made clear tangible progress in moving over to Wayland and completely shutting off X support with only Xwayland for compatibility then you’re not a systems engineer. You’re a zealot. Every year the BSDs become just a little more irrelevant. Wayland is not a shiny new squirrel. It’s your inevitable future. If you don’t have the time, resources, engineering skill or requisite number of contributors, then it’s time to retire your project. If Wayland is not the default for all BSDs even with X still an option at install by 2026, it’s time for a mass migration away from BSDs. This zealotry nonsense needs to be squashed.

I don't see why Wayland needs to be the inevitable future for BSDs. Like you, I prefer things progressing and evolving; that is why I use Linux, not BSD. I use Wayland not X11. But if NetBSD wants to conserve their original forms and protocols, they are naturally free to do it and they deserve respect for their choice. It is the same for a Linux distribution that wants to keep either X11 or Wayland. I think X11 fans trying to stop others using Wayland or Wayland fans trying to do the other way around are both zealotry. In the end, developers do what they want, users choose what they want. Things get eliminated or promoted, we move on.

[roelandjansen]

So far, one of the blocking issues I have with wayland is that copy paste in a vmware workstation linux vm doesn't work to and from the underlying windows host.

Also nomachine doesn't work, black screen.

Both can be a vmware and a nomachine issue but for the user like me it is what it is. Wayland: does not do what I need; xorg works.

[billyswong]

OK, this is a good one, although I never used such eyecandy (but not stating that's useless because I don't use it). However I was pretty convinced that this "decoration madness" (rotating cubes, etc.) died out counless years ago.

TBH I'm a bit sad that openbox won't work on wayland (I have my "standardized" .config/openbox dir carried around from machine to machine in the last ~15 years) but I think that's the fault of openbox devs and not the X11 > Wayland switch and I've moved on.

The rotating cube is mostly eye candy animation. The roll-up of a window into just the titlebar serves some functionality as an alternative form of minimize. It is not eye candy. The widespread of CSD from GNOME conflict with such design though.

[Britoid]

I don't see why Wayland needs to be the inevitable future for BSDs. Like you, I prefer things progressing and evolving; that is why I use Linux, not BSD. I use Wayland not X11. But if NetBSD wants to conserve their original forms and protocols, they are naturally free to do it and they deserve respect for their choice. It is the same for a Linux distribution that wants to keep either X11 or Wayland. I think X11 fans trying to stop others using Wayland or Wayland fans trying to do the other way around are both zealotry. In the end, developers do what they want, users choose what they want. Things get eliminated or promoted, we move on.

That's fine but they'll have to come to the realisation that eventually toolkits and applications will drop Xorg support and it will be up to them to either maintain said applications or adopt the Wayland protocol.

[oiaohm]

So far, one of the blocking issues I have with wayland is that copy paste in a vmware workstation linux vm doesn't work to and from the underlying windows host.

Also nomachine doesn't work, black screen.

Both can be a vmware and a nomachine issue but for the user like me it is what it is. Wayland: does not do what I need; xorg works.

https://kb.nomachine.com/AR02P00969

roelandjansen; nomachine with Wayland has different instructions. Black screen with nomachine is normally not using the instructions for Wayland and end up in Xwayland. Yes the limitations are also included in the above link at this time.

It is possible to make nomachine connected to Xwayland somewhat work with XWayland Video Bridge but the other way is less CPU and more stable..

I mostly use RDP off weston when I need remote.

[Britoid]

Where exactly is the evidence for this? You have evidence that a piece of software cannot be improved? Could it not be the same behaviors and beliefs that are bogging down Wayland development were also applicable to X11 development?

What exactly are the edge cases that require us to deliberately push out X11 - and why can't separate programs be used on top or beside X11 to remedy them?

X11 is a drawing API and doesn't aline to how modern graphics hardware works with surfaces and textures.

We need the latter for things such as HDR, hardware compositing (zero input lag), zero copy composition, tile renderers (popular in mobile GPUs). Why do you think Android doesn't use Xorg but instead used a display protocol similar to Wayland (with far less function)

This would be a huge change to the Xorg way it doing things so they decided to just have a clean cut implementation and put Xorg through a compatibility layer.

Microsoft and Apple have both changed their protocols over time but have the advantage in that they control the system toolkits which every application links to, thus can update apps by updating those, we don't have the same luxury in Linux hence why this has taken 2 decades.

[mrg666]

That's fine but they'll have to come to the realisation that eventually toolkits and applications will drop Xorg support and it will be up to them to either maintain said applications or adopt the Wayland protocol.

Well, this is for them to worry about, right? They always catch up to complain about the next step of progress. That is how it is, I guess.

[kpedersen]

That's fine but they'll have to come to the realisation that eventually toolkits and applications will drop Xorg support and it will be up to them to either maintain said applications or adopt the Wayland protocol.

I hope the two (BSD and Linux) go separate ways in more ways than this. For too long Linux has kind of been using Xorg because there was nothing developed in-house. As such it has added a little bit of mess to the way Xorg is developed (which Xenocara is one attempt to fix, xbin is another)

Yes, BSD communities might have to maintain an X11 backend for i.e Gtk in a few decades, likewise Linux communities will have to develop and maintain a SSH/Wayland+Pipewire backend for OpenSSH if they wan't a secure remote graphics solution.
Honestly, this isn't really a big task. At work we have an SDL2 backend for Gtk2 which was written by only a few guys. GUI is easy.

They *are* different operating systems, so this extra work is to be expected if project goals are to be correctly maintained.

It gets to the point where if everyone was too scared to implement stuff, Linux wouldn't have taken off and everyone would still be talking about Windows right? BSD communities are even larger now (and growing) than Linux was during 2010. Yes, probably because there are more open-source developers around now than back then; but its fun that this fact doesn't change anything.