Re: sysvinit vs. new powerd...
You (Ian Jackson) wrote:
> Michael Alan Dorman writes ("sysvinit vs. new powerd..."):
> ...
> > 2) remove powerd from sysvinit entirely, and make the new powerd
> > conflict with older versions. Easy.
> ...
> > 3) make powerd replace sysvinit powerd. Also not unreasonable, but
> > upgrading sysvinit could do unexpected things.
> ...
> > 4) use update-alternatives to allow selecting one. I don't know
> > exactly what this entails.
>
> I have no knowledge about powerd as a specific application, but when
> you've decided what the final arrangement of packages should be
> *please* come and ask me *before* releasing any new packages with
> funky things in the control files.
>
> If you tell me what the state of the packages is now and what you want
> it to be I'll tell you what to put in the control files.
>
> If you don't ask me you'll probably tie yourself in knots.
I was thinking about releasing a sysvinit without the powerd included
(nobody I know uses the vanilla version anyway). This would be
sysvinit-2.63-1. Nothing would change in the sysvinit control file.
I do not plan to put a Recommends: powerd in the sysvinit control file.
The powerd control file would then have something like:
Package: foo-powerd
Provides: powerd
Depends: sysvinit (>= 2.63-1)
This would work since sysvinit is an essential package and already installed
anyway. The only thing that would happen if a seperate powerd package is
installed that dpkg forces one to upgrade sysvinit as well, which is
exactly what we want. Right?
Mike.
--
Miquel van | Cistron Internet Services -- Alphen aan den Rijn.
Smoorenburg, | mailto:info@cistron.nl http://www.cistron.nl/
miquels@het.net | Tel: +31-172-419445 (Voice) 430979 (Fax) 442580 (Data)
Reply to: