Comment by cf100clunk

2 days ago

In what way was Bruce incorrect, your one link excepted?

he is counting every c file in the systemd _repository_ which houses multiple projects, libraries and daemons. he equates that to the c file count for a single init. it's a disingenuous comparison. systemd-init is a small slice of the code in the systemd repository.

  • I'm guessing he shares my belief that systemd-init cannot exist in the wild on its own, correct? When you want a teacup, you have to get the whole 12 place dinner set.

    • IIRC the mandatory components are the init system, udev, dbus, and journald. Journald is probably the most otherwise-optional feeling one (udev and dbus are both pretty critical for anything linux regardless), though you can put it into a passthrough mode so you don't have to deal with its log format if you don't want. Everything else is optional.

      1 reply →

  • > he is counting every c file in the systemd _repository_ which houses multiple projects, libraries and daemons. he equates that to the c file count for a single init. it's a disingenuous comparison.

    See, this is why when I refer to the Systemd Project, I spell it as "SystemD", and when I'm referring to systemd(1), I spell it "systemd". I understand that some folks who only wish to shit on the Systemd Project also spell it that way, but I ain't one of them.

    > systemd-init is a small slice of the code in the systemd repository.

    Given the context:

       Yes, systemd provides a lot of capabilities, but we will be losing some things I consider important.
    

    I'd say that the topic of discussion was SystemD, rather than systemd. systemd doesn't provide you with all that many capabilities; it's really not much more than what you get with OpenRC + a supervisor (either supervise-daemon or s6).