[Pkg-sysvinit-devel] Bug#851747: sysvinit-utils: unmaintained package should not be Essential

[Simon McVittie]

Control: retitle 851747 sysvinit-utils: drop Essential flag
Control: severity 851747 wishlist

On Wed, 18 Jan 2017 at 23:21:46 +0000, Ian Jackson wrote:
> I have done this now, with a changelog entry that closes these two
> bugs.  I think that is appropriate but I'm sure someone will complain
> if not.

I still think it would be good for sysvinit-utils to not need to
be Essential, because the Essential set is larger than it needs to be.
(In particular, minimal/buildd sid chroots as produced by debootstrap
don't have /sbin/init any more, because a chroot doesn't need it;
so it seems weird for them to be required to contain helpers for
sysvinit/LSB init scripts that they will never run.)

So I would prefer if you didn't close #851747, but I'm dropping its
severity: a maintained package in the Essential set is a lot better
than an unmaintained one, but a smaller Essential in buster would
be better still. This is clearly not something that will/should be
fixed in stretch though.

There is some discussion on
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=810018 about what
remains in sysvinit-utils. While I don't think pidof really deserves to
be Essential, it's still too widely-used to not be; but other
distributions ship pidof from procps, and we could do the same. The rest
of the entry points (e.g. fstab-decode, with 2 users in the archive)
are sufficiently narrow that their users could likely just gain Depends
instead.

This would mean that the (newly revitalised) sysvinit packaging team
would only need to be responsible for systems that actually boot using
sysvinit, and not systems that boot using systemd or init-less chroots
that don't "boot" at all.

    S