To be more precise, Gnome is actually not dependent on system or logind, it is dependent on the DBUS services that logind provides. So all it needs to run Gnome with all features supported is a project willing to write software that provides these services.

1 members found this post helpful.

Not sure what is meant by "let Mate in" but while Debian has supposedly chosen Xfce for the default, for now, Mate is actually in the Jessie repo or will be when released. I haven't checked the testing repo for it but Mate is in the Sid repo.

What about all the other crap like Seats and the like, iirc it needed those too, along with some other stuff like the rest of the user-session-units or however that was called, also can power management be done with D-Bus only or does it require other components since iirc that was another one of the things it was about to break.
Really, systemd is a bloated octopus, wanna bet more and more stuff will get even more strongly coupled within it.
Also IIRC a few of the DMs also had to be rewritten because of systemd and it also ate consolekit, not that I care that much though I can just have an init script spawn few gettys on several of the ttys and if I want to go to GUI I will simply modify my bashrc to start it if I am in a particular runlevel and on a particular termina;


Good, I should really start using testing or unstable, rather than stable one of these days.

I think if it was less of an octopus, and more or this quiet little thing that just exists in the system and isn't wrapping tentacles around every program, it might have a better favor towards people seeing it, but that's part of the problem.

The fact that it brings in and requires non-critical components that SystemV never did, nor does any other init system, is a big put off. Why do we need d-bus so badly? Why does ConsoleKit need to be abandoned? Did udev really have to be rolled in as a part of systemd? Why do we need logs written in developer languages? To me, all these questions lead a good argument that overall, sometimes some things are just inclined to be optional and not required.

All this stuff is provided over DBUS by logind.
No, consolekit ate itself. It was maintained by a Gnome developer, but since they changed to logind that developer has stepped down from maintaining it. There is no maintainer for it for quite some time (git-log) and despite so many people complaining about consolekit being deprecated no one stepped up and took over the project.

1 members found this post helpful.

Would be a good and noble venture if developers here at LQ or another community could step up and take over the project even if temporary. A lot of projects could be easily revived if someone would just take up the reigns.

Interesting change of events:
https://lists.debian.org/debian-ctte.../msg00281.html

If I'm reading that right, and if I'm not then so be it, but he seems to be wanting to simply ignore Ian's proposal and pull a full coup-de-tat against him and the Classic Debian design frame by forcing an issue rather than further studying the long term effects of this.

Not to be negative but I see yet another distribution choosing poorly, rather than wisely.

As I see it, he merely wants to decouple the question for the default init system from the question how dependencies on init systems are managed after they have found a new init system.

In other words he is putting the Cart before the Horse, IMO the right choice of action is to decide if you want the extended init dependencies first, then decide which init system to use and how to fork/modify it.

As it seems, the votes are now in and, though not officially declared yet, systemd has won. However, Ian Jackson, one member of the CTTE, is extremely angry about the process for the last voting and wants to dethrone the chairman of the CTTE. This seems to become like the plot of a soap opera.
I am not happy with the decision to go for systemd, but seeing now the state in which the CTTE is further diminishes my trust in Debian.
For me the decision to migrate of my servers away from Debian wherever possible is finalized.

So even if systemd is the default, can it be replaced with sysvinit since both are in the repository? Or will sysvinit completely go away?

If systemd becomes the default, you'd have to remove and restore a LOT of packages as systemd becomes a main system dependency. Anything built on top of systemd would theoretically have to be reinstalled against a non-systemd-ified package. That's a lot of work, and it would make maintaining sysvinit less desirable to the team, and it would open a doorway to say, "dump it" if it came down to it by vote.

At this point... I'm already learning enough about FreeBSD and PC-BSD to make an educated choice if this hammer of death falls Linux-wise.

Most likely it will go away. That's the whole point of changing the default init, after all. Though in case of systemd win it might be interesting to see how nonlinux ports will cope with it...

Same here. Although in the past I haven't had the greatest luck with hardware under FreeBSD.