MX Linux Forum

asinoro wrote: Sun Jul 28, 2024 12:32 pm

anticapitalista wrote: Sun Jul 28, 2024 12:17 pm I disagree with this point.
The vast majority of software is init agnostic - it simply does not care what init is running.
The problem is that Debian packagers add systemd/libsystemd0 as a hard dependency even if the app does not require it at all

Since you are one of last Mohicans that fight the beast, what other alternatives are away of Debian which could be user-friendly?

MX has never claimed to be free of systemd so as long as their sytemd-shim keeps on working, there is no issue at all and MX-24/25 won't be any different under the hood as present or previous versions. ie it offers sysVinit and systemd boot on one iso.

FWIW I doubt wayland will work very well without systemd by the time wayland is pervasive. Personally I already am finding a number of linux components aging out (pretty much like me). I have moved to using systemd and find it actually seems to be more stable on Debian (given the stuff I use). But keep in mind, nothing moves very fast in the Debian world. But Linux will eventually move from x11 to wayland and when it does things will be different (not to be confused with better or with worse either for that matter).

@manyroads ... Linux will probably move from x11 to wayland (and a lot of DE and WMs) , but an other init system will not be, and must not be, an obstacle.

oops wrote: Sun Jul 28, 2024 2:59 pm @manyroads ... Linux will probably move from x11 to wayland (and a lot of DE and WMs) , but an other init system will not be, and must not be, an obstacle.

I guarantee you it is and will be a pain. It already is. I gathered some info that may help get you started in solving the situation for yourself...

Wayland Items to follow

If you use these DEs here is where they are headed:
-- KDE Plasma 6: Wayland by Default: https://news.itsfoss.com/kde-plasma-6-dev/
-- Xfce 4.20 Aiming For Usable Wayland Support While Maintaining X11 Compatibility: https://www.phoronix.com/news/Xfce-4.20-Wayland-Roadmap
-- Gnome Wayand is used by default in Debian 10: https://wiki.debian.org/Wayland

Does Wayland require systemd?
-- Here's a response/ discussion from the BSD world. https://forums.freebsd.org/threads/kde- ... emd.77776/
-- You may use internet search and find more, what I have gathered from reading discussions is that there are ways to make wayland work on other inits such as OpenRC but that they are not always trivial or stable.

manyroads wrote: Sun Jul 28, 2024 2:43 pm FWIW I doubt wayland will work very well without systemd by the time wayland is pervasive.

Wayland is a protocol and essentially an IPC specification for a very specific use case (graphics, display and window management), so it doesn't have anything to do with systemd. That is like saying, Gopher won't work very well without Solaris.
There are a several compositors such as Wayfire and Labwc that have nothing to do with systemd.

manyroads wrote: Sun Jul 28, 2024 3:18 pmI guarantee you it is and will be a pain. It already is.

That depends on your device. In my case, switching KDE to the Wayland session solved a lot of problems, such as input lag, screen tearing and high CPU usage. So Wayland has the potential to solve a lot of issues that X11 has, but it's up to the compositors.

AK-47 wrote: Sun Jul 28, 2024 4:45 pm There are a several compositors such as Wayfire and Labwc that have nothing to do with systemd.
I was unclear above. What I was referring to is that there are many tools that run on DEs and wms that will not have x11 equivalents on Wayland and those that do offer Wayland support will those which work and require systemd. Things like conky have a history of exhibiting problems (though they are now largely fixed, at least when I last used conky on Wayland.

As for Labwc and Wayfire, they are both very tough to use. I found them rough and low in functionality (again last I used them).

manyroads wrote: Sun Jul 28, 2024 3:18 pmI guarantee you it is and will be a pain. It already is.

That depends on your device. In my case, switching KDE to the Wayland session solved a lot of problems, such as input lag, screen tearing and high CPU usage. So Wayland has the potential to solve a lot of issues that X11 has, but it's up to the compositors.

Yes KDE is going to work just fine, they have a lot of resources devoted to the transition.

AK-47 wrote: Sun Jul 28, 2024 4:45 pm That depends on your device. In my case, switching KDE to the Wayland session solved a lot of problems, such as input lag, screen tearing and high CPU usage. So Wayland has the potential to solve a lot of issues that X11 has, but it's up to the compositors.

Kwin has a history of lag and CPU utilization on X.org, which can often be seen on low-end hardware. I haven't noticed the same thing with other window managers on X.org. X.org even worked well back when computers were single core. Screen tearing might be related to X.org, but I haven't really seen that either. From a gaming/framerate perspective it depends on game, driver and compositor, but generally I believe X.org and Wayland are comparable in performance at least when I looked at phoronix.com last time. No doubt Wayland is the future for desktop Linux since this is what is being worked on.

The Wayland discussion is similar to the systemd discussion. One may like one or the other. Ultimately, the "industry" (X.org, Red Hat, Valve etc.) decided that systemd and Wayland are the future. This is probably how it will turn out in the long run for mainstream desktop Linux.

I like SysVinit and X.org, but they are becoming more niche. I also think KDE is the future, while smaller desktops might disappear or become more niche. In the end it's manpower that decides and every deviation from "mainstream" has a cost associated with it. From a casual user (application user) perspective the underlying tech doesn't matter that much.

On the other hand the vast scope of MX Linux (live/installed, SysVinit/systemd, multiple DEs and now X.org/Wayland) might have been what brought MX Linux to Distrowatch number 1 and kept it there.

I see systemd as dependency hell in the init world. One service automatically starts another etc. However, if the MX live system can be brought to systemd I personally don't think SysVinit is that important. On the other hand SysVinit was the core reason I started using MX Linux (systemd didn't give me a good feeling). Systemd works well on MX Linux and I could switch to it without much thought.

Since I see myself as something in-between a casual user (application user) and a tinkerer I don't really know what foot to stand on. For tinkerers SysVinit might be crucial, otherwise soul is lost so to speak. These days I mostly see desktop Linux as an application platform.

Things like Wayland (compositors), desktop environments (KDE, Gnome, Enlightenment etc.), Pipewire and Flatpaks could tip the scales in favor of systemd. If you don't need "big software" then several systemd-free options exist.

@dreamer ... SysVinit versus Systemd , and X.org versus Wayland , are two separate and independent things. (Wayland is more secure for the fishing than Xorg)

One of the reasons for choosing of MX Linux was its use of SysVInit. Nothing against systemd. However systemd has some very valid concerns raised against its use. As far as moving to systemd, it will become inevitable by the time this decade is up. For example it is almost impossible for GNOME to work without systemd. Even KDE has certain applications, like KDE System Monitor, which do not work fully with SysVInit. This will increase as we go ahead.

Additionally even systemd documentation says that it will stop supporting SysVInit scripts somewhere down the future.

Debian might have had other init systems and support for them in the past, but it is no longer the case. And it will no longer be supported in the future. Most of the significant financial support for Debian comes from people running Debian on servers. These guys would prefer systemd and not SysVInit. Also we should not overlook the power of conformity. For maintainers, having systemd controlling the bootup and services means having less of scripts to be maintained and tested per release. That is a very powerfull pull for distros. Do not underestimate it.

Like the wise Borg said, it is futile to resist. All will be assimilated.

LinuxSpring1 wrote: Wed Jul 31, 2024 7:55 am As far as moving to systemd, it will become inevitable by the time this decade is up.

What a sad and defeatist attitude to have about the future.