Shared publicly  - 
 
At last, the official statement of what's going to happen with upstart and systemd in Ubuntu 12.10 and beyond. Even though it is not the outcome I was favoring, I'm happy that we have a decision now, as well as an official stanza to point upstreams to. This is much better than the previous situation where I had to reply to upstreams based on conjectures and personal preferences, and it was hard to do any longer-term planning.

So, thanks +Steve Langasek for making the announcement!
8
Martin Pitt's profile photoTomeu Vizoso's profile photo
3 comments
 
It's a good explanation. What's that different outcome you mentioned?
 
From my perspective of a plumbing stack maintainer/packager I would have preferred going systemd all the way. With upstart, Ubuntu needs to maintain the old ConsoleKit/udev-acl/GNOME fallbacks by itself, and we cannot benefit from the already existing systemd jobs that are being included in various upstreams and distros, nor cooperate with them. Also, while udev still is a standalone thing and works with upstart right now, there is no guarantee that this stays so forever. It is my understanding that Ubuntu wants to compete at the UI level, not really on the plumbing/kernel one, so following Fedora/SUSE on their plumbing path would seem like the path of least resistance and effort to me. In fact, with my systemd PPA I have a booting and running Ubuntu 12.04 with less than 6 hours of effort.

However, I appreciate that in the server world the requirements are quite different, and there is no (proven) existing systemd integration anywhere yet. I honestly cannot judge how much effort this would be. The foundations team considered both sides here, and got to the conclusion that the migration effort and risk on the server side outweighs the extra effort on the desktop side, so I respect that.
 
Yeah, well, next release, right? :)
Add a comment...