GhostBSD Makes XLibre Its Default Display Server

Info:


View: https://x.com/Ericbsd/status/2014513591433474392

I don't have an opinion. Have fun commenting. (I don't know why I write things like this. The megalomaniac in me, I guess).

Edit (pasted from my comment below):

BTW, I believe it's forum user baaz who the main developer of GhostBSD thanks for their port.

 
Last edited:
Being an almost complete ignorant, I like the philosophy of the X server much more than that of Wayland. The X server let's the advanced user do a lot of cool things like the well known possibility of executing a graphical app on the server and getting the interface on your "workstation" (I hope I'm not saying any falsehood).

I am using Wayland, though, because X11 is slower in my computer and I get screen tear. These things don't happen with Wayland. I wonder if XLibre will make X11 better in these aspects too.

Edit: Just to provide more info, I don't use Nvidia, my GPU is Intel. I've never gamed. I just watch YouTube and other streamings.
 
The X server let's the advanced user do a lot of cool things like the well known possibility of executing a graphical app on the server and getting the interface on your "workstation" (I hope I'm not saying any falsehood).
True. We used to do that. The other nice hardware option was to have a big powerful server, with multiple displays (+keyboard +mouse) connected, so several people could use the same machine for graphical work. That's one of the reasons that the X displays are usually called ":0.0"; the second person might use ":1.0" and so on.

In practice, I haven't seen any of this in use in a decade or two. Today, physical displays are connected via very high bandwidth interfaces to the GPU of exactly one computer, which in turn is connected via a high bandwidth interface to memory and CPU. Putting a network in between is just not done any longer.

There is one exception: For IT service, sometimes consultants will take over the screen, but there screen bandwidth does not matter.
 
True. We used to do that. The other nice hardware option was to have a big powerful server, with multiple displays (+keyboard +mouse) connected, so several people could use the same machine for graphical work. That's one of the reasons that the X displays are usually called ":0.0"; the second person might use ":1.0" and so on.

In practice, I haven't seen any of this in use in a decade or two. Today, physical displays are connected via very high bandwidth interfaces to the GPU of exactly one computer, which in turn is connected via a high bandwidth interface to memory and CPU. Putting a network in between is just not done any longer.

There is one exception: For IT service, sometimes consultants will take over the screen, but there screen bandwidth does not matter.
I hate speaking about Windows, but the server feature that I think is called WinApps (I'm not sure) is precisely apps running on the server and displaying on the users' machines. I installed this system in my company. It was pretty cool because you could have a very powerful server running CPU-consuming apps concurrently and the result was faster than running them locally on average computers. It was a local server, but with today's internet speed, it could very well have not been.
 
I hate speaking about Windows, but the server feature that I think is called WinApps (I'm not sure) is precisely apps running on the server and displaying on the users' machines
Sounds cool but I'd be more confident in seeing everything already with RDP :p

Can't imagine doing it on *nix, but doing nginx conf with GUI floating folders, notepads, and a small Firefox window was more convenient than messy in the moment :p (I'd update a conf, save, refresh on Firefox, then update the notes on Firefox; all-in-one same machine vs floating SSH on dedicated client doing things to server)
 
BTW, I believe it's forum user baaz who the main developer of GhostBSD thanks for their port.

 
As nvidia doesn't state supports for XLibre (and as far as I could heard of, unlikely at least for now), if any issue specific for XLibre (means, cannot reproduce on Xorg) on nvidia GPUs are needed to be fixed / worked around by XLibre side, at least the issues are in binary only part.

If the issue is caused by something that XLibre side fixed mis-behaviors (here, different as documented by Xorg, means, Xorg's fault), Xorg should cherry pick the fix from XLibre and nvidia should chase the changes.

If the issue is in any parts provided as source codes, we (nvidia driver ports maintaining team) can dig into and try fixing, IF THE FIXES ARE NOT AT ALL HARMFUL TO XORG. But it would take time, as we're volunteers with limited time.
 
Back
Top