On Tue, 11 Oct 2016, Christian Hofstaedtler wrote:
> * Santiago Vila <sanv...@unex.es> [161011 01:35]:
> > As of today, lsb-base is "less essential" than before, because
> > util-linux no longer depends on it in stretch.
> >
> > Also, if I'm not mistaken, lsb-base is a shell library to be used by
> > init.d scripts, but those scripts are obsoleted by systemd.
>
> They are obsolete when replaced by systemd native service files, AND
> we are no longer bound to support non-systemd init systems. (Maybe I
> have missed a change to that?)
No, you have not missed anything, but some packages check for existence
before using functions from lsb-base, as a way to avoid a dependency,
so in many cases we can avoid a "hard" dependency on it.
> As long as this is not the case, all packages that ship a daemon will
> have to depend on lsb-base, or reimplement the log_* functions PLUS the
> systemctl redirect logic.
>
> It could have been argued that init-system-helpers ("helper tools
> for all init systems") should depend on lsb-base...
>
> How does this change improve the situation, when it is implicitly
> forbidden for a system to have lsb-base missing?
Sorry, I don't understand the question.
A system without lsb-base is not "forbidden", implicitly or otherwise.
Maybe it was "implicitly forbidden" when util-linux (an essential package)
had a dependency on it, but as I said, that's no longer the case.
In my particular setup (building packages) I don't even need an init
system at all.
Thanks.
