Wayland Clients Can Now Survive Qt Wayland Crashes / Compositor Restarts

[Barnacle]

We are still in the "early days" of Wayland honestly. Discovering the problems and trying to find the best solutions. At least every single application doesn't have to implement this functionality, just a handful of toolkits.

A core development team that constantly says "that's your problem, not ours, because we say it's out of scope!" is not "trying to find the best solutions". The fact that they are forcing every DE and UI toolkit to reinvent the wheel endlessly is not "trying to find the best solutions". Refusing to fix the original broken standard after it has proven itself to have grave shortcomings is not "trying to find the best solutions". Attempting to force developers to radically reconsider how graphical applications should work in order to support a platform that has well under 1% of the desktop market is not "trying to find the best solutions".

One for be forgiven for thinking some great Corporate Enemy of Linux has "sponsored" the core Wayland developers, because everything they do is completely counterproductive, fruitless and wastes the time of the entire open source, cross-platform desktop application community. But yeah, everybody keep holding out hope that they can get it right eventually, they will just need another 13 years to do it.

[Myownfriend]

One should remember that design choices in the OSS world are oftentimes held at the same level as sacred commandments received from God(s).

We have another one.

[smitty3268]

One for be forgiven for thinking some great Corporate Enemy of Linux has "sponsored" the core Wayland developers

That's actually the core of the problem right there, believe it or not.

Nobody is willing to sponsor core wayland or X developers. That's why they decided to have the minimal amount of scope there, and it's why anything that needs to get into wayland takes so long, because there's no manpower there to do it. Just like why getting something like HDR into X has taken forever, because no one is paying people to make it happen.

Ideally, of course, there would be resources put in place to help with that, but 20 years of X development showed that basically no one was willing to work on such code without being paid, and basically no one was willing to pay people to work on such code.

Like it or not, there's a lot more ability to fund people to work on Kwin or Mutter, because those are tied much more closely with distros and how their users interact with the system, so that's where they decided to dump all the hard work to.

Over time, no doubt best practices will be discovered and the various compositors will work towards consolidating shared code. Just like how over time everyone came together in X.org and dropped all the other duplicated code in other implementations. But we're still a long way from that now. Lots of people still have different opinions about how things should work and they're all choosing to try their own stuff rather than compromise on some perceived lesser solution.

[NeoMorpheus]

This is massive, this is welcomed, but God damn it I don't understand why each Wayland toolkit has to implement all these features again and again.

Why can't we have something akin to I don't know Win32 (xlib for X11) which all other toolkits can build upon and not reinvent the wheel all the f-ing time? Something which draws fonts, primitives, etc. etc. and is being accelerated at the same time, e.g. using OpenGL (ES), Vulkan, etc.? It's sad to see so much human resources being wasted.

And on the other hand we have absolutely the same situation in regard to Wayland compositors which reimplement a metric ton of features again and again. A nice neat idea of having/painting/presenting perfect frames (all good, I like it) turned into a nightmarish implementation scenario. I'm just rending the air, dismiss this comment as lunacy. The lunacy which doesn't exist in well established commercial OS'es such as Windows, MacOS, iOS, Android or even QNX.

This is one of the reasons I said for Valve to consider (back in the day) to use HaikuOS as the base for SteamOS 3 and what I meant when I said "politics of Linux". Clearly, my lexicon is not as advanced to properly express it as done by you.

Again, was just a crazy idea to explore, but still a good one i think.

[Nozo]

This is one of the reasons I gave for Valve to consider back in the day to use HaikuOS as the base for SteamOS 3 and what I meant when I said "politics of Linux". Clearly, my lexicon is not as advanced to properly express it as done by you.

Again, was just a crazy idea to explore, but still a good one i think.

It wouldn't have been very smart of Valve to use an eternal (barely) beta as a foundation, if they wanted to use a permissively licensed OS and make a proprietary SteamOS 3 and the Steam Deck a closed platform, there's the old, stable, reliable and (unfortunately) parasitized FreeBSD.
And if you don't believe it, then look at Sony and its Orbis OS.

[Quackdoc]

Nice! hope this can go a long way towards live gpu swapping. being able to swap the render GPU live is a great feature for KVM, eGPUs and apparently waydroid

[NeoMorpheus]

It wouldn't have been very smart of Valve to use an eternal (barely) beta as a foundation, if they wanted to use a permissively licensed OS and make a proprietary SteamOS 3 and the Steam Deck a closed platform, there's the old, stable, reliable and (unfortunately) parasitized FreeBSD.
And if you don't believe it, then look at Sony and its Orbis OS.

I am going by how capable BeOS was and how good KaikuOS is given the little amount of developers involved.

I am simply looking at something that could had been great now, if Valve poured millions of dollars and manpower as they have done with Linux but in HaikuOS instead, a couple of years ago.

Or maybe, as you pointed out, Sony should had approached Valve with AMD and do a partnership to bring Orbis to PC's and hit MS (and hopefully intel and nvidia) where it hurts.

Its simply a nice What If scenario that I would like to think would had been great.

[acobar]

I am going by how capable BeOS was and how good KaikuOS is given the little amount of developers involved.

I am simply looking at something that could had been great now, if Valve poured millions of dollars and manpower as they have done with Linux but in HaikuOS instead, a couple of years ago.

Or maybe, as you pointed out, Sony should had approached Valve with AMD and do a partnership to bring Orbis to PC's and hit MS (and hopefully intel and nvidia) where it hurts.

Its simply a nice What If scenario that I would like to think would had been great.

The reason people use Windows is not because it is Windows, it is because of many applications that run best on it or even only on it. Autocad, Adobe Suite, MS Office, and on and on. The money Valve could put on any other OS would not modify this simple fact.

[alphabitserial]

Except that X11 is very stable right now and KDE Wayland is not, and we really don't know yet how long it will take to mature. Also, I will not even start to comment on the atrocious state of remote desktops under Wayland.

Cheers.

I'm using KDE Wayland, on an NVIDIA system, every day for work. Every day I remote in to Windows systems to administer them. After work, I use it to play my games, without worrying about them being Linux native/running in Wine/native Wayland/XWayland. I have hardware accelerated video playback working in both Firefox and Chromium. There are still some small papercut issues remaining for sure, but it's quite usable and a better experience for me, otherwise I would not be using it.

[ferrellsl]

We are still in the "early days" of Wayland honestly. Discovering the problems and trying to find the best solutions. At least every single application doesn't have to implement this functionality, just a handful of toolkits.

You're effing delusional if you think we're still in Wayland's "early days". Wayland's initial release was 30 Sept. 2008. It's now 2023 in case you've lost your calendar.