to clarify:
The developers of the Alpine Linux-based postmarketOS mobile distribution today that they’re now supporting the systemd init system alongside OpenRC and other alternative init systems.
and:
postmarketOS currently supports the Sxmo, Phosh, GNOME Shell on Mobile, and KDE Plasma Mobile UIs. While the Sxmo images will stay with OpenRC, the GNOME and KDE Plasma Mobile images will be built on top of systemd
thank christ, it sucks to see projects let perfection be the enemy of good.
a project can be as ideologically pure as it wants, but it’s not gonna accomplish much if that handicaps it out of anyone seriously using it.
Plasma mobile looks way better than phosh IMO. Glad that the mobile environment will resemble the desktop environment. Makes it easier to do cross-platform stuff.
Phosh kinda sucks, idk how it is on Librem devices, but on PmOS it’s super clunky.
Dunno what it looks like now, but back then it sucked on Librem too. I can understand trying to do different stuff, but they made the experience worse and I haven’t looked back from KDE mobile.
Great, this should help attract the mainstream linux user more
PostmarketOS has had really strange priorities lately. I’m not a fan of the whole ethos of Ubuntu mobile (including their use of SystemD) but at least they have stuck to actually getting every feature working on some devices with reasonable specs. My computer uses KDE and OpenRC and has far fewer issues than it did on SystemD. This feels like a waste of resources to reinvent the wheel.
Reinventing the wheel is what they were doing without Systemd.
On their announcement they cite various instances of having to write polyfills and ending up with basically ‘Systemd at home’ but buggier.Reinventing the wheel is what they were doing without Systemd.
Weird. We had those wheels before Lennart’s cancer showed up.
No, init systems sucked before. It was a bunch of poorly documented and poorly managed shell scripts
We had this fight for years, your side lost. Evolve or die.
I honestly don’t know why you were downvoted so much. You could have get very different responses in a different forum.
Personally, I scoff at anyone who discredits Poettering’s contributions.
He actually did things for the Linux ecosystem, unlike a lot of people on these forums.
Not a big fan of people who sit on their hands and only get off of them to tell others to sit on theirs.
Now, Microsoft employee. And everybody is OK with that. New generations suck.
The project is in an too early phase to debate over SystemD. Can you guys please hold back with these arguments until pmOS reaches at least 4% market share.
There is no minimum market share threshold to discuss the way the software you use is being developed and PostmarketOS will not reach 4% in the foreseeable future (and it probably never will). Desktop Linux only just reached that threshold after decades of work and systemd arguments have been happening for years regardless. The conditions for mobile Linux are considerably less favorable. If we can’t discuss systemd until 4% is reached, we can’t discuss systemd ever. Which is fair, because the systemd horse has already been beaten to death at this point. But not because it hasn’t reached some arbitrary 4% threshold. That makes no sense.
If we can’t discuss systemd until 4% is reached, we can’t discuss systemd ever. Which is fair, because the systemd horse has already been beaten to death at this point.
Exactly :)
Well you’re right but the more postmarketOS grows, the harder it is to switch to another init system
They are giving options, no one is forced anything. People should complain upstream at init systems and
desktopmobile environments.Why would you want to. Systemd is the standard for a reason.
It does have disadvantages. The only real advantage of it is the completeness of system administration tools. Since they aren’t that much needed on a phone and the performance of that class of devices is not groundbreaking, using another init system is a good idea. Though it depends on what the specific user wants of course. As long as there is a way to change the init system, it shouldn’t be a problem
Another init will be slower and will require much more time and resources though.
Don’t believe. Do you have any proof of that?
Systemd was created to allow parallel initialization, which other init systems lacked. If you want proof that one processor core is slower than one + n, you don’t need to compare init systems to do that.
Systemd is the standard for a reason.
- bad build process
- ignoring best practice
- RedHat forcing it on the planet
- people forgetting that every deliverable of systemd is a lie.
Lmao
I don’t have an opinion on the whole systemd debate but are you going to expand on what you’re meaning, or will just keep spewing bs bullet points? Specially n4, wtf do you mean by that?
Can you explain in a little more detail?
And? :)
ive been working on migrating away from systemd myself, so much headaches. I like the services setup, but man the issues can sometimes be baffling
Isn’t Linux without systemd just a hobbyists niche exercise in masturbation though, let’s be real.
I mean, I don’t really care what it is so long as it works fine.
I’ve been becoming more accustomed to using systemD and seeing it as an integral part of my system.
Replacing systemD to me seems about as asinine as replacing ‘cd’ or ‘print’ at this point. Why bother?
I simply need something less … self integrated to work with. With systemd it can be a real pain when you need to do something different with a lot of things. Take for example udev. It’s neat, everything relies on it now, but systemd’s implementation, the official implementation, is kinda trash.
If you for instance run a distro in a chroot (I realize this may be niche to most people, but its critical for me) systemd’s udev is useless. So you replace it with eudev or in my case libudev-zero, All is hunky dory except that in nearly every distro, libudev is directly mixed with systemd init, if you use systemd, you use libudev.
And yes, This is indeed fundementally a packaging issue, and ofc I can rebuild/repackage systemd to not provide udev, I’ve done it before, and probably will again. But these are the kind of issues I wind up running into and so many more. If it was a one off kinda thing then it would be fine, annoying but fine, but sadly for me it just isn’t.
I do actually like systemd as an init system, I just wish it didn’t do everything else, At one point I had high hopes for rustysd a service manager capable of running systemd services but development seems to have stopped.
