John Hughes <john@???> writes:
> On 18/12/15 15:50, Mitt Green wrote:
>>> It's a library whose sole purpose is to make sure that
>>> packages *don't* depend on
>>> systemd.
>> So, you are saying that libsystemd0 is harmless and it
>> doesn't mean anything unless you install systemd, systemd-sysv and so on?
>
> Exactly.

*If* libsystemd0 is structured as you claimed it was (I didn't check
this), it's nothing but a second (and presumably somewhat sub-par)
implementation of a set of systemd defined APIs which sits in a shared
library (very likely written by the systemd developers, as it's part of
the same source package as systemd) and some runtime switching code
which either uses the facilities of identifiable systemd-programs or
falls back to the library reimplementation in case these aren't
available. It's purpose would be to enable applications depending on
systemd-facilities to work on systems which apparently don't have it
installed by providing a second systemd implementation in another file,
IOW, the 'preferred programming interface' of such a system is whatever
the systemd developers decided it should be and this interface is
implemented by code which is part of systemd.

But when the systemd developers effectively control both the API
definition and the API implementation, the question whether or not such
a system is actually running systemd becomes largely academic: It's a
systemd system, just possibly one using a different way of managing
processes.