Re: On init in Debian
Russ Allbery, le Fri 30 Mar 2012 10:41:05 -0700, a écrit :
> Samuel Thibault <firstname.lastname@example.org> writes:
> > Russ Allbery, le Thu 29 Mar 2012 23:41:40 -0700, a écrit :
> >> It's focused on being clean, supportable, and fully integrated with
> >> Linux capabilities, *not* to solving everyone's use case, even to the
> >> detriment of being universal.
> > So that directly conflicts with making it a default init implementation.
> On Linux? Why?
> It's not trying to be a *standard*,
It is apparently trying to be a *Linux* standard, being adopted by all
distributions. That does mean things which, even if not talking about
unix but just Linux, means you have to take some care, in the same vein
as when you define a Unix standard.
> But I don't think that being a standard in that way is a horribly
> compelling feature that Debian cares about. It is sort of nice to have
> everything use the same init script format, and it used to be that was
> kind of, sort of the case, but it's not been true for years.
The current init standard boils down to /etc/init.d/foo start/stop, and
it has been true for years. The particular content being another matter.
I'm not saying it was a good standard. The deviation of the content of
the init scripts really is a matter.
> For better or worse, there is no standard for init scripts, and neither
> systemd nor upstart (nor, for that matter, sysvinit) are really trying to
> become that.
If they are to be adopted widely, it'd be better for them to sort of
being one, so that upstreams could ship configuration snippets, instead
of seeing all distributions defining its own ones, bringing small
discrepancies here and there, which can be a pain when going from one to