Not trying to start a flame war or anything.
I'm honestly curious how systemd "creeping features" compare with alternatives, e.g. the machinectl shell and systemd-boot.
I have experience with networkd and it seems fine, but I don't really have an in-depth basis for comparison.
I'm also curious whether the criticisms in http://blog.darknedgy.net/technology/2015/10/11/0/ are important and/or can be resolved.
no tinfoil pls
In general the systemd components are better. They're modern, leverage powerful building blocks from the rest of systemd, and learn from the mistakes of previous attempts at those tools.
The issue that non-trolls have with systemd is that we'll grow too dependent on these currently good tools and at some point past that they stop being good and start being bloated and we have no way to drop them without dropping everything
Most of the extremely long writeups, especially anything that mentions "binary logging" or "non-modular", are generally not worth the time to read and are written by people who know very little about systemd because they refuse to use it (or use it exclusively in the way that their preferred distro ships it and are entirely unaware of upstream options)
>>55719454
>The issue that non-trolls have with systemd is that we'll grow too dependent on these currently good tools and at some point past that they stop being good and start being bloated and we have no way to drop them without dropping everything
Can't they just be refactored/replaced? "We might have to fix it in 5 years" doesn't seem that bad...
>Most of the extremely long writeups, especially anything that mentions "binary logging" or "non-modular", are generally not worth the time to read and are written by people who know very little about systemd because they refuse to use it (or use it exclusively in the way that their preferred distro ships it and are entirely unaware of upstream options)
I think the writeup I linked to is the opposite of that.
>>55719549
>Can't they just be refactored/replaced?
By a fork, sure. Good luck just suddenly switching all the infrastructure and users to a less bloated fork if systemd starts getting too fat.
In reality it's probably not an issue. The systemd developers are pretty competent and the quality of their software that's convinced almost every distro to rapidly switch to it shows that. But it's what some critics are worried about regardless.
>I think the writeup I linked to is the opposite of that.
The writeup you linked is notoriously not the opposite of that, and is actually exactly that but with a ton of disclaimers and promises that it totally isn't that. The arguments in that article all boil down to:
- Complaints about non-issues
- Complaints about downstream configurations (not systemd issues)
- Appeals to philosophical puritanism that have evidence against them and aren't causing problems
All wrapped up in a bunch of fluff and raised hands
>>55719672
>Good luck just suddenly switching all the infrastructure
The code isn't modular/OO enough to to do it piecemeal?
>The writeup you linked is notoriously not the opposite of that,
really? I've noticed some boot problems on my system, specifically getting to GDM but it not being responsive to input. Is this not a possible side effect of non-deterministic booting?
>>55720026
>The code isn't modular/OO enough to to do it piecemeal?
Try reading more than a few words into a sentence
>Is this not a possible side effect of non-deterministic booting?
No