Live data from Hacker News

Wayland is not ready as a 1:1 compatible Xorg replacement just yet

gist.github.com

361–370 of 419 posts

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#361

Earlier quoted context omitted.

It's not just setting a new property. Meanwhile, today, you have clients that are dpi aware and are not setting the property. And they all will be out in the wild for an undetermined amount of time (because stuff works for the application developers, why bother?). If you do not handle them anyway, your users will complain that your desktop is broken.

It is just setting a new property. I am not sure we're on the same page here, but if your existing client is DPI-aware, and you just want to let the compositor know it, then it is a one-line change in your client. As for the "Meanwhile" part, again, how is that better with a new display stack? Unless the point here is that it will be better to just break all those (DPI-aware) clients unconditionally, rather than actu…

Because that "meanwhile" takes years. There are still clients that use libX11 ("it works for them") instead of libxcb. libxcb was introduced in 2003.

So "meanwhile" you would have apps that are not DPI aware without flag, DPI aware without flag and DPI aware with flag. And even crappier desktop as you have now for years to come.

The new API is not just for DPI; it is for many reasons, DPI is only one of the issues.

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#362

Earlier quoted context omitted.

I specifically qualified it with efficiently . As in, when the surface is in the GPU memory, the software doesn't have to read it over PCIe bus and then put it back into another buffer for the encoder; but that the encoder can do zero-copy encoding in the GPU memory. Here, Firefox uses dma-buf. Unless X11 exposes a special extension for exposing dma-buf, your client won't be able to do that.

You said it yourself. I even believe that extension already exists since it must be going on inside DRI3 for some EGL/VK extensions to work right now. Don't think it worth to create an entirely new display server API for just this. And in practice, XShm-using OBS actually works better right now than this specific apparently Gnome-only dma-buf API. https://www.youtube.com/watch?v=kD70ur_xTmE

dma-buf is linux-specific, not Gnome specific. It is buffer management on GPUs; it is the low-level API that Mesa and libva, and yes, DRI too, build on top.

Just because some specific application works better (for some values of better) on the Xshm does not say anything. In year or two, the situation can be very different, permanently.

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#363

Earlier quoted context omitted.

> How can you complain about Ubuntu LTS keeping Xorg as default in an LTS when RedHat itself had Xorg as default at that time. Isn't this hypocrisy or irony? RHEL8 uses Wayland as default. RHEL8 was released in May 2019. Ubuntu LTS was released in April 2020. What hypocrisy you are talking about? > (same thing of idiots complaining about upstart being shit while RedHat and Goole were still using and supporting it. Re…

So did you check the show stopper bugs that Ubuntu found? or they should ignore the bugs to please RH and it's fanbase ? Maybe the next LTS could profit by the fact that RH recently was forced to fix the bugs in Wayland (though not sure how many paying customers started using it yet and submit tickets for the devs to fix)

Show stopper for Ubuntu was screen capture. So screen capture works today, and if they switched, if would be difficult only for the first few weeks, when Ubuntu is buggy anyway.

But they didn't and the long-term consequences of delaying it will be felt across all linux distributions for years to come. So much for appeasing RH and it's fanbase.

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#364
post #19

> Wayland solves no issues I have This is a common sentiment, but it is kind of funny. Because software engineers working for companies will know the struggle of trying to justify reducing technical debt, improving security practices, etc. but fail to see why these things are useful to them personally when the tables turn. Wayland indeed “breaks a lot of shit” but it’s not an accident or done due to incompetence. Als…

There's no 'winning' this discussion. The users clearly want faster horses.

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#365
post #30

I really miss the network transparency in Wayland. Just being able to start a program on another machine through an SSH session is so useful. You can use something like vnc but a "whole desktop in a window" is slower and not as handy. I'm on freebsd mostly anyway which isn't big on the Wayland train so I'm fine for now. But it would be great if X11 development would be continued.

> You can use something like vnc but a "whole desktop in a window" is slower Really? I’ve found x-forwarding much slower and annoying unless the host is on a LAN.

Agreed, my workflow is a workspace with firefox with jira gitlab etc. open. Then a few with terminals and vscode open, several of those are often vscode running on remote machines via x forwarding. Then I develop on one box that is a couple vpn hops away and its too slow for forwarding. I hate it though because I have to switch to that workspace, fullscreen the window, and then when I'm done, esc the vnc and switch to another workspace. If there is a wayland way to have just firefox or just vscode running remotely that would solve my concern.

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#366

Earlier quoted context omitted.

So did you check the show stopper bugs that Ubuntu found? or they should ignore the bugs to please RH and it's fanbase ? Maybe the next LTS could profit by the fact that RH recently was forced to fix the bugs in Wayland (though not sure how many paying customers started using it yet and submit tickets for the devs to fix)

Show stopper for Ubuntu was screen capture. So screen capture works today, and if they switched, if would be difficult only for the first few weeks, when Ubuntu is buggy anyway. But they didn't and the long-term consequences of delaying it will be felt across all linux distributions for years to come. So much for appeasing RH and it's fanbase.

I disagree, this happened before with pulseaudio, it was not ready and for years people would recommend to remove it, so the leason is not replace something that works now with something that is shiny but might work next year.

I also think NVIDIA had a role in this default display issue, Ubuntu has to use something that actually works on all hardware combinations,competent users can install and try a Wayland implementation or all of them.

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#367
post #84

Earlier quoted context omitted.

They aren't though. And X is pretty deprecated by now.

> They aren't though. Remote desktop, game screen recording, screenshots, how do those work?

As I said, there are pain points. If you need application X to do Y and it doesn't work in wayland that might be a dealbreaker for you.

I'm having no issues with screen recording and screenshots. Haven't attempted nor need remote desktop but know people that do it. Obviously don't have the same breadth of alternatives such as X.

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#368
post #13

Not much constructive here... I'm of the opposite opinion. Think twice before doing anything with X as the experience and knowledge you gain will soon be useless. Instead invest in wayland, it works great and will only get better and X will only get worse at this point. The pain points are few but yes, if you must have screen sharing with a particular app that doesn't work with wayland that might be a dealbreaker. If…

Something that isn't particularly good 12 years in will likely never be so. At 12 years in some really basic stuff is just now rolling in. In general the idea of natural progress is completely imaginary. The general nature of things is to get worse not better entropy always wins.

X isn't particularly good either, and even older ...

Re: Wayland is not ready as a 1:1 compatible Xorg replacement just yet

#370
post #345

Earlier quoted context omitted.

You have a better chance of being hit by a meteor than falling victim to a compromise that would have been stopped had you only used wayland. Also xwayland is not a fix for any of the issues identified.

That's not remotely accurate, but thanks for writing!

Ok let's address the problems of which you assert 99% can be fixed via xwayland.

First problem identified

> Wayland breaks screen recording applications

Can xwayland fix it? Nope.

> Wayland breaks screen sharing applications

Can xwayland fix it? Nope.

There is very little Linux desktop malware in the wild.

What there is would seem to be minimally effected by wayland example

https://animajav.us/linux-systems-hiddenwasp-malware-trojan/

Instead of worrying about trying to get the keyboard as a non root user logically one might attack the filesystem or just work on getting root.

Linux desktop security vs threats actually already running on the system appear to be more an asperational goal than a reality.

At least there are documented cases of meteorites injuring people I can't find Linux malware available in the wild that would be mitigated by wayland.

Post reply on HN