Bug#727708: logind working without systemd
Thomas Goirand <zigo@debian.org> writes:
> Sun, 19 Jan 2014 13:42:40 +0200, Russ Allbery
>> Hopefully, logind will continue to work without systemd and people will
>> volunteer to maintain the necessary packaging for that configuration,
>> and none of this will be a problem.
> I really wish you were right Russ. Because that's not what upstream is
> doing (since systemd 205, it's not the case), and Debian package
> maintainers have stated this as an argument in the favor of systemd.
> I would very much like the tech ctte to express itself on this topic on
> the final statement (whatever default init system is chosen).
Here's the problem with doing that: what, exactly, do you think the TC can
say, given our powers and given the hard rule in Debian that no one is
going to be forced to do work they don't want to do?
Here are some possibilities that would arguably be within the remit of the
Technical Committee:
* The lack of a logind that works without systemd as PID 1 is an RC bug in
systemd, so systemd will be removed from the archive if it doesn't
provide this feature. This is just silly to say, since removing it from
the archive doesn't provide a logind that works without systemd, so this
doesn't achieve anything useful.
* NMUs to revert systemd to version 205 or earlier are permitted if the
systemd maintainers try to upload a newer version where logind doesn't
work without systemd. This is clearly not a defensible solution. We
can't hold the package indefinitely at a back revision and lose security
support, etc. It might make some sense if the hard dependency of logind
on systemd were a bug that would be fixed in a later version, but that
doesn't appear to be how upstream views the issue.
* NMUs to fork logind to work without systemd are permitted and the
systemd maintainers may not revert them. *Assuming* there is some
irreconcilable conflict between the systemd maintainers and the people
who want to do this work (and I don't think that's a correct
assumption), I don't understand why we would take this approach as
opposed to packaging a non-systemd implementation of logind separately
with the required Conflicts and so on, so that the individual
maintainers can work on things that they care about without having
constant tension over NMUs.
* The systemd maintainers are required to cooperate with people who are
doing the work to make logind work without systemd so that they aren't
prevented from maintaining those packages. I see no reason why the TC
should say this given that there's no sign that the systemd maintainers
would not do this voluntarily. Obviously, *right now*, there are
reasons to wait to see how this whole discussion will resolve since
that's going to significantly change the priorities and shape of this
work, but I see no reason to believe people wouldn't be able to work
this out in some reasonable way without the TC involvement if people are
available to do this work.
I don't see the point in the TC saying any of those things, and I think
some of them are actively destructive. Furthermore, I think all of them
are very ethically questionable. If I were the systemd maintainers, most
of these statements would strike me as heavy-handed extortion: an attempt
of the TC to work around section 2.1.1 of the constitution and force me to
do work by holding something else I care about hostage unless I do that
work.
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: