The implicit premise of this comment is that linux is broken and needs to be changed. It isn't. The changes are not inherently good. Development is not inherently good. Just look at Gtk3 from 2014 to 2024. It was far more functional in 2014 (re: keyboard input) and now that has been removed because "progress".
Yes, theoretically we could just stick with old software that was perfectly good at the time. The problem is that the ecosystem moves together - there's an emergent consensus about what is "current" (i.e. what's obligatory to support) across many different projects.
So each component is interrelated, and holding one back means sacrificing compatibility with the new features (and security fixes!) of many other components.
In this way we can find ourselves dragged into using software that is actually worse than it used to be. This is important to note because it means our use of a component is not proof that it's good. It just means that the ecosystem is good enough overall that it can force us to accept devolved versions of certain components.
It makes sense to argue about what the consensus of the ecosystem should be even while recognizing that we will probably accept it regardless.
My issue is not with systemd. My issue is with the argument that all development is good. In this example I am pointing out how Gtk3 has suffered from development attention from GNOME over the last decade and became worse.
Maybe run0 is worse than sudo. Maybe not. I have no personal experience on that topic and I doubt anyone here does.
Reality is more nuanced than that, because even if the newer software is worse, if it markets itself better it can persuade other projects to drop support for the old software.
Indeed. I have been using Ubuntu 14.04 under ESM for the last 10 years. But that ESM support is ending in 2024. I've tried modern distros using modern Gtk3 and they're lousy with file chooser bugs. I've tried patching Gtk3 gtkfilechooserwidget.c myself but I can only fix it for the first file->open dialog, not subsequent ones. Attempts to get help in #Gtk over the last 5 years have been rebuffed in IRC and ignored/closed repeatedly on the issue trackers: https://gitlab.gnome.org/GNOME/gtk/-/issues/5872
We don't have to use their stuff. systemd actually does a good job of managing services' up and down state - it's the rest of the bundled functionality that people take issue with - and if service up and down is all you want, there's runit.
Wayland seems to be a solution without a problem, and it's only winning by default. I've toyed with the idea of forking Xorg - the code is a bit odd by today's standards, but I didn't find any direct problem with it, and it works fine - the biggest problem I discovered in Xorg is a lack of project management as nobody knows what it should and shouldn't do
My gentoo system has some gentoo-specific fork of udev without systemd in it.
I guess but in this case it's doing something that sudo doesn't do — get privs from a daemon instead of setuid, and using something that systemd already did (systemd-run).
Is it that weird that they would make a runas frontend when all the pieces of it were lying around?