On 8/15/23 20:17, Darryl Gregorash wrote:
On 2023-08-15 19:17, bent fender wrote:
Seeing systemd as a distinct threat to the kernel itself I'll stick with those who are for a choice, but I'm not that interested in convincing anyone of anything.How is systemd a threat to the kernel?
Of course I can't speak for Mr Fender, but I think I know what he was referring to.
It's perceived by many that systemd is in violation of the UNIX Philosophy, where programs should do only one thing, but do that one thing very well. Wikipedia has an extensive description:
https://en.wikipedia.org/wiki/Unix_philosophy Such people must have felt greatly threatened when such packages as LibreOffice started showing up on the scene. There is absolutely no relationship between a word processor and a DBMS, yet LO and all others
On 2023-08-15 23:22, Lew Wolfgang wrote: like it can do both, and plenty more besides.
systemd strives to incorporate many functions into one binary blob. I'm
not saying this is necessarily bad, but it's not UNIX. The concern then arises that with an expanding systemd it will encroach farther into kernel territory, that carried to its logical extreme will completely push out the kernel. systemd then becomes the operating system, the thought of which concerns many. More nonsense. Again, a cursory inspection of the systemd control files shows nothing to suggest there is any encroachment whatsoever into kernel territory. The oS developers here are certainly in a far better
Nonsense, which even the most cursory inspection of any systemd control file will show (the systemd control files are all stored in /usr/lib/systemd and sub-folders). All systemd does is to ensure that everything is started in an orderly fashion, that no service gets started until all of its requirements have been met. For example, if a service must not be started before the network is fully configured, it will contain a "requires" or "after" statement to that effect. The service itself is started by the appropriate script or executable included in the software package. For example, here is the control file for D-Bus: [Unit] Description=D-Bus System Message Bus Documentation=man:dbus-daemon(1) Requires=dbus.socket RefuseManualStart=true RefuseManualStop=true [Service] ExecStart=/usr/bin/dbus-daemon --system --address=systemd: --nofork --nopidfile --systemd-activation --syslog-only ExecReload=/usr/bin/dbus-send --print-reply --system --type=method_call --dest=org.freedesktop.DBus / org.freedesktop.DBus.ReloadConfig OOMScoreAdjust=-900 The D-Bus service requires a listening socket, so the daemon cannot be started until that socket is present. Only then does systemd load the daemon that actually does all the work. position than I am to address such issues, but IMO, any suggestion that systemd is somehow "out to replace" the kernel is nothing but FUD.