![](https://seccdn.libravatar.org/avatar/cabdbf4d350ab6a15265803acab1634d.jpg?s=120&d=mm&r=g)
John Andersen said the following on 08/05/2013 02:47 PM:
You think I'm missing the point and have no idea what my point is, then you proceed to directly answer the question about the autogeneration based on init.d scrips being a kludge.
(Is it against your religion to answer a direct question without first preceding it with an insult?)
So: I think you're missing my point and I think that I'm missing your point? How is saying so an insult?
Without some indication by opensuse as to their plans for the future, building custom systemd scripts is likely to cause as many problems as it will solve.
That's a perfectly valid view but I'm not sure we are going to get an answer to that. In the mean time we need to get what we have working. In my case that mean customizing units. In the future I expect it will continue to mean customization as the 'on size fits all' isn't going to fit me, nor do I think it will fit many others. In Istvan's case getting Samba to work right now has, I suspect, a higher priority than what might or might not be in some future release. If Istvan understands what is going on with systemd & samba as I've tried to explain, and perhaps if Felix sends him the F19 unit files, and he sees how they fit in and replaces the /etc/init.d scripts, then that knowledge will equip he to deal with some hypothetical but as yet unstated plan in a future openSuse release. Compared to scripting, the declarative form of the unit files is amazingly simple.
Why not fix the sources used by the generators in the current version(s) and wait for formal release of the non-generator version of OS. (Aren't you the one always berating Linda for the degree to which she customizes her system?)
The reason for not fixing the generators is that replacing the relevant scripts with .service unit files is quicker, easier, more stable and is the direction things are going - as evidenced by F19. It doesn't require knowledge of C programming, compiling etc. KISS and all that. Unlike Linda's customization this is in keeping with the direction of evolution. And as I say, it is a simplification.
Buy the wya, Your own source suggests generators are ok in his pet case: "Generators should only be used to generate unit files, not any other kind of configuration."
Once again my opinion is that you don't get it. Or perhaps you're simply being argumentative. "Pet case"? Conversion of a table to unit files - /etc/fstab or the ttys - is much more straight forward, true, but the whole point is that generating anything except unit files isn't going to be of much use at this point in the boot sequence. In an hypothetical system we could do without /etc/init.d, /etc/fstab and much else and have everything predefined as unit files. I don't know about your opinion of such a configuration but I'd hate that. There are many things that are easier to maintain as text tables such as /etc/fstab than as a long list of separate files. There are principles of keeping like things together. Yes, I'm aware of the change from the single file table of the old InetD to the multiple files of XInetD. That made me wonder, but at least they were few in number. Compare # systemctl -a | wc -l 250 And many of those are auto-generated. I'd hate t have to maintain all those as discrete files by hand. With the auto-generator many of them they are created, used and disposed of. Auto-generators have their place, but using them to deal with items in /etc/init.d that are yet to be converted is a kludge, and one that we should overcome. -- How long did the whining go on when KDE2 went on KDE3? The only universal constant is change. If a species can not adapt it goes extinct. That's the law of the universe, adapt or die. -- Billie Walsh, May 18 2013 -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse+owner@opensuse.org