[systemd-devel] RFC: removing initctl support
Lennart Poettering
lennart at poettering.net
Thu Sep 24 07:05:47 PDT 2015
On Thu, 24.09.15 15:51, Tomasz Torcz (tomek at pipebreaker.pl) wrote:
> On Thu, Sep 24, 2015 at 03:01:21PM +0200, Lennart Poettering wrote:
> > That stackexchange link lists a pile of garbage. We have an official
> > API to check whether the system is booted with systemd:
> > sd_booted(). It's documented here:
> >
> > http://www.freedesktop.org/software/systemd/man/sd_booted.html
> >
> > And we even document on that man page what precisely it does
> > internally (which is equivalent to access("/run/systemd/system/",
> > F_OK) >= 0) and suggest people to reimplement that simple check in the
> > language of their choice, even in shell... That way, they don't even
> > have to link against libsystemd.
>
> And then there is this sabotage:
>
> „This check is already broken, because uselessd creates this directory too”
>
> uselessd% git grep 'mkdir.*/run/systemd/system'
> src/core/main-no-init.c: mkdir_label("/run/systemd/system", 0755);
> src/core/mount-setup.c: mkdir_label("/run/systemd/system", 0755);
> src/core/unit.c: mkdir_p("/run/systemd/system", 0755);
>
> Meh.
Why would anyone care? It's not that anyone was actually using
this... Also, by doing this, that software explicitls *wants* to be
treated like systemd, hence why not let it?
If people focus on weird init systems like that, that's certainly
their own fault...
Lennart
--
Lennart Poettering, Red Hat
More information about the systemd-devel
mailing list