Wayland's Weston Is In Severe Need Of More Development Help

[cipri]

Well, that's kind of the point, as Wayland is a protocol

Yes, naive like this I was too! Till working on my own wayland client, because libwayland-client is crap, and later I understood I was not possible to get rid of libwayland-client (you always had to link against libwayland-client)!
It seems now recently with vulkan 1.1 one maybe might find a solution.

[M@yeulC]

Nobody seems to be addressing the fact that all of this wayland(in general) gnashing of teeth is pointless if Nvidia is NEVER going to release a driver for it. Their only concern and focus on playing nice with Linux is machine learning. All of you Wayland numbnuts can GFY.

Until that very fk ass large majority of the GPU's on planet Earth are addressed, Wayland is a pipe dream for hippies and dipshits.

With all due respect, I think you are grossly overestimating the marketshare of nVidia GPUs. Various sources (here's one) quote the Intel iGPU marketshare at around 70%. The rest is probably split between AMD APUs, GPUs, and nVidia GPUs. If you take no AMD APUs, and go by the GamingOnLinux statistics of nVidia 70% -- AMD 30%, you can expect a ~20% marketshare for nVidia, which is a conservative estimate.

Going by this, the kwin maintainers' decision to first target Intel iGPUs seems pretty reasonable

And if nVidia's only concern is to play nice with machine learning, I don't see why people that are concerned with something else than machine learning would pick nVidia. It actually seems that AMD's marketshare is on the rise, both GPU and CPU-wise, and for good reason.

cipri You are probably right, I should try to write a wayland client one of these days, and judge by myself.

[jpg44]

Everyone is implementing its own Wayland implementation... This is something what we have not seen in X and it's disturbing to me.

Yes I have mentioned this before. I have said before Wayland is ridiculous waste of time reinventing the wheel when we already have X, and where X did things the right way. Wayland took all of X's good design decisions and decided to throw them all away and do things in the most horrible ways possible. Gone are the display server/window manager seperation, mechanism not policy, common Display Server implementation as a result of WM seperation, not including window manager and look/feel code in the display server, network transparency, application/hardware isolation due to the stream based protocol, etc. With Wayland hardware drivers are inside applications. Hard to think thats smart.

[jpg44]

Nobody said that mir is better , I for myself dont know Mir in Detail, but I know libwayland-client and weston are exemplary pieces of crap. Should go in the hall of fame: how to write a piece of crap in 3 steps.
Because of that crap that people struggled with, it took like 10 years and wayland is still not widely adopted. Who wants to contribute to crap? Crap I can write myself, I dont need to start learning and struggling to understand the crap of others. Go check the official pdf of the wayland documentation.... WTF, WTF, WTF, WTF.... was my reaction. It was a documentation that doesn't allow you to write to simplest wayland client. Should have been called fairy-tale, not documentation.

Wayland i have thought is a waste of time and not well designed. Awaste of programmer time and resources. It came from the same GNOME people that gave you the piece of crap that is Gnome 3. Take a look at Gnome 3. Enough said. The niave hubris, self-righteousness and elitism of this group, and their disdain, contempt for the users, for you, they could care less what you think, is an insult.

Its a waste of time: because Linux has long had a perfectly fine UI stack with X11. All wayland has done is waste programmer time reinventing the wheel on a poorly designed display stack that is actually FAR LESS flexible than x11 because: NO network transparency, NO WM/Server seperation, among many other reasons.

The time would be better spent working on something Linux doesnt have yet, maybe, better hardware support, 100% Windows WINE compatibility, decent voice recognition or OCR, perhaps. I mean COME ON, this is why Linux still sucks on the desktop, and falls flat regarding a number of kinds of applications such WYSIWYG content creation.

These wayland people are asking people to waste their time on Wayland which just badly reimplements something we already have with X11 when I just listed many things Linux does not have yet. Wayland actually has set back Linux development.

Gnome 3 is such a brain damaged piece of nonsense that the whole branch should just be deleted and we should go back to Gnome 1, seriously. The retardedness of Gnome 3 and the waste of time that is Wayland is ridiculous work of GNOME 3 team.

[M@yeulC]

With Wayland hardware drivers are inside applications. Hard to think thats smart.

Didn't we debunk this in another thread the other day? Hardware is used the same way under X and Wayland. Only the way results are displayed changes slightly (and then, not so much because of all the new X extensions). The Wayland API should just be much simpler, and have a lot less round-trips.

[duby229]

Everyone is implementing its own Wayland implementation... This is something what we have not seen in X and it's disturbing to me.

It's an interface, you have to make your own implementation of it... I mean after all it works just like any other interface....

[jpg44]

Didn't we debunk this in another thread the other day? Hardware is used the same way under X and Wayland. Only the way results are displayed changes slightly (and then, not so much because of all the new X extensions). The Wayland API should just be much simpler, and have a lot less round-trips.

X Window System has always had a network transparent stream for apps to send graphics to the server. Both the Core X protocol and later on GLX extension for 3D graphics. DRI was added way back in the early 2000s as an *Alternative* rendering path by placing the hardware drivers in an X application. But the idea was at that time, users could also continue using GLX protocol where GL commands are sent over the X protocol to the server and then can be sent to drivers. This was called Accelerated Indirect GLX. So you should be able to choose to use GLX or use DRI, depending on your preferences. DRI is not necessarily bad, but people with a security concern or who need network transparency should have the option of using GLX.

X Window System does need a better fine grained security model for what clients should be able to access which resources and state in the display server. A normal client should only be able to access the window state for its own windows, none more. A screenshot or remote desktop, or screen reader, utility could be given more permissions that it needs.

The X Window system has needed some work on the internals to move to OpenGL rendering backends and to implement AIGLX as well, only keeping XAA/EXA for legacy drivers.

Visual artifacts can be solved AFAIK with Double Buffering extensions and Vertical Synchronization extensions.

None of this needs a new display server.

[jpg44]

Wayland might as well be called Microsoft Windows. Its everything we disliked about Windows, on Linux, It takes away the best feature, network transparency. If I wanted a OS where I could not choose which display an app should be displayed to and is this monolithic thing that takes away all configuration and choice., I would use Windows. GNOME 3 is such an obvious copy of Windows and actually , its worse, Windows has more configuration than GNOME 3 will give you. Gnome 3 is the worst UI I have ever seen, it really is atrocious. and it treats the user with contempt, you cant configure anything.

[molecule-eye]

Wow, the Wayland haters are out in force today
I personally can't wait to get rid of X! Which I'll do once kwin + plasma will have agreed upon a protocol for the virtual desktop pager, and activities (though that's less important to me).

Both of those are very low priority, especially because virtual desktops work, it's just the pager widget that doesn't. I'm more concerned about how difficult it is to get clipboard history functionality, to get scrolling to be as smooth as as on X with libinput, and to get the big browsers to run natively on wayland (which is why I'm currently using Falkon--decent, but no firefox).

[jpg44]

Both of those are very low priority, especially because virtual desktops work, it's just the pager widget that doesn't. I'm more concerned about how difficult it is to get clipboard history functionality, to get scrolling to be as smooth as as on X with libinput, and to get the big browsers to run natively on wayland (which is why I'm currently using Falkon--decent, but no firefox).

With all due respect, thats a lot of work, for what, your just duplicating functionality already available on Linux for ages. Why not do something we don't have. Like: Decent OCR support, accurate voice recognition, an open source natural language translator, WYSIWG web design tools, better WINE WIndows API support, Support fot the Presentation API for .NET in Mono.

Whats the point in redoing the same features over and over again?

What does wayland solve? Nothing. It does nothing we don't have with X or could not get by improving X.