/g/ — Is systemd really that bad?

>>109057969 (OP)
I think there's dinit which is closer to systemd but less comprehensive, but I'm an idiot on these topics. i use openrc on distro, not for any fancy reason just coz i feel like it

yes
/thread

I love s6, it's dead simple in a good way

>>109057969 (OP)
Yes. Worse is effectively not having init freedom, and systemd et alia active work to to reduce init freedom.

>>109057969 (OP)
I just recently switched to linux, systemd was default and I don't know yet why should I care when everything works.

systemd is just better than the trash that came before it and after it. The only problem you can lob at it is the fact everything is getting tied to the later systems like logind which will make untethering it for a successor init system much more difficult.

Look at this thread. You NPCs really still dont understand why we are against systemd. Are you really this thick?

>>109057969 (OP)
>what's stopping me from putting in a reverse patch in /etc/portage/patches/ and calling it a day?
Yourself, because you won't do it.

>>109057969 (OP)
Yes because its based off of launchd, goes into territory far beyond what an init system should touch, and is so annoying to avoid that most non-systemd systems still have to use parts of systemd for compatibility. Also, more and more things are becoming reliant on systemd so Gnome can no longer get ported to the BSDs or non-systemd Linux distros without basically forking it and KDE seems to be heading in the same direction (and people seems to have amnesia about how Gnome did the exact same thing, so people are massively coping about). The fact that you need to fork a distro and change make your own repos to use a non-systemd version of your distro should be explanation enough as to why its a problem.

Its a solution to a fake issue. Linux being fragmented isn't a weakness, its literally the reason why its so successful. You can break off and make your own distro whenever you want and you still get to passively benefit from the greater Linux ecosystem. Meanwhile if you fork a BSD, you basically either become a glorified custom ISO, fall behind and become irrelevant, or you effectively become an unrelated OS. NetBSD and OpenBSD can't share projects and code willynilly despite OpenBSD being a fork, Dragonfly is far behind FreeBSD and basically on life support despite being a fork, and GhostBSD is just a custom ISO for niggers to lazy to set up their own FreeBSD.

Systemd is just one of the many examples of Red Hat and co. wanting to control every aspect of Linux (and turn it into Mac 2.0), so that all the other distros need to follow their lead or get left behind to mingle with the Gentoos and BSDs of the world. Considering Fedora is Red Hat Linux Beta, anything that happens there you can 100% guarantee will happen in Debian, Ubuntu, and Arch 0.5-2 years later. You'll eventually have to fork the entirety of Linux, which will be far too much work for a handful of autists.

>>109059532
Debian adopted systemd because it was overall the best available option. If Canonical wasn't so retarded with its Contributor License Agreement, maybe Debian and Ubuntu would be running Upstart today.

>>109057969 (OP)
It's not satan like half of the board will claim, but it's pretty bad.

The init system itself is fine, only turbo autists complain about that. Most of the processes are retarded, but you can often just swap them out for sane alternatives.

The worst thing about it is journald, which is a fucking retard faggot service. Anyone who prefers binary logs over plaintext should be executed on the spot.

It is terrible from a philosophical perspective though, as it completely goes against the unix philosophy, plus why would you want red hat controlling large portions of your computer.

>>109060551
Are binary formats like ELF and tar against the unix philosophy too?

>>109059281
All the jew conspiracies in the world don't mean shit because systemd is LGPL at the end of the day.

>>109060551
>journald
so disable binary logging and pipe it to rsyslogd
>unix philosophy
>red hat control
there is no better alternative that meets the needs of serious distros

>>109057969 (OP)
you tell me, have you ever ran into a situation where you thought to yourself
>"systemd is shitting itself again! this piece of junk just gets in my way!"
or does it just work?

you have to realize that most people in this space are ideology schizos. they form very strong opinions based on how things are done or should be done in theory, when in practice it makes effectively zero difference. and i don't think systemd is perfect, or even all that amzing, but not once have i ran into any issues actually using it

>>109059281
>"other people are npcs, but not me. i am smart. why are npcs so stupid?"

>>109060872
>or does it just work?
i think some people are just sensitive to things they perceive as clutter. it's like if you take a box full of stuff and shove it in a closet. one person might think, this is useful, i just don't need to interact with it right now, another might think, this is clutter, this is stressful, if i'm not using these things, then it needs to go away permanently
i think the mindset is similar on software

>>109057969 (OP)
>Is systemd really that bad?
It has some pointless annoyances and stupid design decisions. But at this point it's been the default for so long there's no point fighting it.
>>109059193
The problem is that Debian adopted Red Hat's SystemD with no real pushback or good attempt to make an alternative work. After that, there was no non-contrarian choice. You were going to use systemd or you were basically fucked and relegated to the ghetto. SystemD did not win out by being the best and most popular choice, it won out through the weight of Red Hat's influence.

>>109060587
>waaa yu r le conspircy therist
Grow up autismo. The world is full of shady shit.

>>109057969 (OP)
https://github.com/davmac314/dinit

>>109062079
Notice how I didn't say "theory"

>>109060872
You are stupid though. You are completely unaware of Microsoft's history and their decades of tactics. Or you dont have the basic pattern recognition needed to recognize these same tactics with systemd, which is typical for an NPC. So are you one?

>>109062110
Fair, since Microsoft employees being in charge of what gets merged into systemd on GitHub is not a theory but a publicly observable fact.