The Linux Thread - The Autist's OS of Choice

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
Is the kernel itself next?
Part of the reason poettering started systemd was because Linus yelled at him for trying to push shitcode to the Linux kernel. He has simply placed his retarded ideas between the kernel and user space instead, creating what now amounts to a partial API wrapper around the kernel.
 
I also like to remind everyone that systemd was pushed insanely agressive on all the usual linux-adjacent outlets and onto distros. At the same time, saying something negative about it, no matter how valid, was explicitly verboten. The sheer agressiveness with which this was done was super weird all by itself. Since the reddit-trier cattle that is a lot of the Linux tech field nowadays does like nothing more than to fall in line and moo with the rest of the herd, (lest anyone finds out they don't understand any of this stuff) that was very effective.
 
I also like to remind everyone that systemd was pushed insanely agressive on all the usual linux-adjacent outlets and onto distros. At the same time, saying something negative about it, no matter how valid, was explicitly verboten. The sheer agressiveness with which this was done was super weird all by itself.
That gets my Glow-dar tingling. Is Poettering a spook?
 
I also like to remind everyone that systemd was pushed insanely agressive on all the usual linux-adjacent outlets and onto distros. At the same time, saying something negative about it, no matter how valid, was explicitly verboten.
That is not my recollection. Every distro which eventually adopted systemd had a long debate beforehand in which fierce opinions were expressed in long mailing list threads. There were the familiar pro and con arguments, and neither view was verboten. As we all know, the pro-systemd side won out in nearly every major distro.

The triumph of systemd was not because of aggressive pushing, but because it solved a problem. The init situation had been unsatisfactory for years, and no one had a good solution. A typical set-up had rc.d directories full of symlinks to scripts with filenames like S25sshd, which were expected to be maintained by distro teams and customized by users. These had grown out of control, and dependencies were nearly impossible to keep track of and enforce. Progress on this problem was very slow, and newer projects such as Upstart only helped a little.

It is also the case that designing an init system is a boring task, a curse of many open source projects. Few volunteers were putting in the work to produce a comprehensive solution. Poettering actually did it.

That's not to say systemd is wonderful or that the criticisms of it are wrong, but, from the point of view of distro maintainers, switching to it was a logical choice. It meant that someone else was doing the hard work of dealing with one of the most persistent and unpleasant burdens of building and maintaining a Linux distro.
 
Systemd causes less disruption than Wayland does. SystemD is a bloated init system but there are alternatives like OpenRC. And at least it works.
Wayland is complex display system and we are forced to implement because "muh modern".
 
At the same time, saying something negative about it, no matter how valid, was explicitly verboten.
Oh really, the shit that nearly caused a civil war in Debian?

As you can see this bug report is certainly raucous support for systemd and the only reason there are thousands of messages here is because they were really excited about it.
 
Oh really, the shit that nearly caused a civil war in Debian?

As you can see this bug report is certainly raucous support for systemd and the only reason there are thousands of messages here is because they were really excited about it.
I don't see any reasonable, fair and balanced criticism of the systemd poblem being allowed there. Here's examples of what that would look like:
Poettring should be raped with a hot poker
Poettring should have his skin flayed off and be dragged over a salt pan
 
The systemd debate wasn't squelched, systemd just is flat out better than what preceded it. The concern back then which has bore out into reality is that too much is getting tethered to systemd as dependency. When systemd was introduced it was dropped quite neatly to the ecosystem which was a testament to the unix philosophy. Replacing the init system should have been a crushingly complex task but for the most part it went on without much pain. So many distros are tethering shit to systemd now such as logind or whatever else meme shit exists that removal of systemd for a successor would not be so easy. The original bill of sale was an init system with journaling attached to it because it needed it and you wanted it, now the mainstream distros are tethering a shitload to it.
 
journaling attached to it because it needed it and you wanted it
No, no I didn't.
About the only thing worse than journald was stuff that logged in tai64n, which meant you had an almost human readable log file but without useful timestamps.
At least journald doesn't pretend to be human readable.
 
If you don't like how every Linux distro that matters in the marketplace now defaults to systemd, you can fix your problem by switching to Windows 11.

I think I'm feeling sufficiently obnoxious to make the point that GCC is so very famous and, yes, so very useful only because Linux was developed.

It's also because Stallman carved out a massive exception in the GPL for GCC headers. If Stallman consistently applied copyleft to his own damn compiler, nobody would use it.
 
If you don't like how every Linux distro that matters in the marketplace now defaults to systemd, you can fix your problem by switching to Windows 11.



It's also because Stallman carved out a massive exception in the GPL for GCC headers. If Stallman consistently applied copyleft to his own damn compiler, nobody would use it.
Or better. Switch to Windows 2000.
 
The systemd debate wasn't squelched, systemd just is flat out better than what preceded it. The concern back then which has bore out into reality is that too much is getting tethered to systemd as dependency. When systemd was introduced it was dropped quite neatly to the ecosystem which was a testament to the unix philosophy. Replacing the init system should have been a crushingly complex task but for the most part it went on without much pain. So many distros are tethering shit to systemd now such as logind or whatever else meme shit exists that removal of systemd for a successor would not be so easy. The original bill of sale was an init system with journaling attached to it because it needed it and you wanted it, now the mainstream distros are tethering a shitload to it.
I regret that I only have one trashcan to give.

An init system starts shit, that's not a complex job. It has nothing to do with syslog.
You can't blame it on the distros when stuff like GNOME had a hard dependency on systemd for a while, and Poettering hasn't done anything to discourage any of this.
 
I regret that I only have one trashcan to give.

An init system starts shit, that's not a complex job. It has nothing to do with syslog.
You can't blame it on the distros when stuff like GNOME had a hard dependency on systemd for a while, and Poettering hasn't done anything to discourage any of this.
Except GNOME is shit too so it's on the distros for making it default. Granted, until Cinnamon came around there weren't a lot of stable options
 
Linus yelled at him for trying to push shitcode to the Linux kernel
I really like the kernel development philosophy of "don't break userspace", and I was actually wondering why Lennart doesn't apply something similar to systemd not breaking user software or systems given how important it is.

An easy example is the famous 2038 issue (1143), where reaching a specific date on 32bit platforms wouldn't let systemd run, so his init system didn't even init, which I'd consider a breaking bug, and so have others from the issue. Some systems had this date set when their RTC battery died.
I think it shouldn't be the case that once we reach the magic day, user space stops booting.
This was a kernel issue, it kept sending infinite notifications causing systemd to busy loop, but these notifications come from a timerfd from systemd.

So if you have an edge case where old systems might not boot because of infinite notifications from a timerfd you created, you probably just stop listening to that timerfd, let the system boot but have lots of errors about it so the user can fix it, right?

No. Lennart doesn't care, it's not his fault. If you can't update your kernel to the latest bleeding version that has this bugfix and your battery dies, you're fucked.

This isn't to say that systemd is the spawn of Satan, I haven't personally had any issues with it, but Lennart is insufferable.

Regarding his run0 stuff, it's nice to be an unprivileged user in case you accidentally do something retarded, and even if an attacker gains access or you run some malware, you should not be able to do any real damage outside of that user's space.

However, if without authentication you can switch to a privileged user or run commands as one, then any actual protection you had goes down the drain. Having sudo/doas/run0 in the first place is the vulnerability. From helping out my buddies use Linux, I see that if a command requires root privileges they just give it root privileges without giving a shit about what it does. The tutorial said to use this command, they are using this fucking command. Look at LTT removing his DE by uninstalling steam, he didn't even need sudo.

If an attacker gets access to a user who can either `doas`, `sudo` or `run0`, you're fucked anyway, so who does it even protect to use sudo instead of just switching to root for your commands? If you do a lot of stuff in the terminal you're likely going to switch to root anyway, because typing sudo/doas/run0 before everything is annoying.
 
Back