Re: Bug#6386: Bug in deb-check
srivasta@datasync.com (Manoj Srivastava) wrote on 02.01.97 in <87ybeckpkv.fsf@tiamat.datasync.com>:
> >>"Chris" == Chris Fearnley <cjf@netaxs.com> writes:
>
> Chris> 'Manoj Srivastava wrote:'
>
> deb-check> bug: usr/sbin/kernel-packageconfig without manpage
> Manoj> Ok, though this is an internal binary. I wrote a man page that
> Manoj> says that.
>
> Chris> Then it should be in /usr/lib, right?
>
> Umm, I see no such directive n the policy or the programers
> guide, or the fsstnd. Also, there are a number of *config{ure}
> programs in /usr/sbin, I was just following what I took to be
> standard practice. I know that gcc binaries live in /usr/lib, so
> standard practice may be debatable; but the point is that nothing
> tells me that /usr/sbin is the wrong location for internal binaries.
Well, my impression was that the various *configure programs _aren't_
"internal binaries". Admins are expected to want to call them explicitely,
to configure some packages.
If kernel-packageconfig is the same, a real man page seems sensible.
Otherwise, it's clearly something very different.
As to internal binaries in /usr/sbin vs. /usr/lib, there seems to be no
clear answer, mostly because it is often hard to decide if a binary really
is internal. The boundarys are quite fuzzy; lots of mainly-internal-use
programs have some mode of interactive invocation: sometimes only for
debugging, sometimes for exotic features, sometimes because one executable
does several different functions (like smail). Just where do we draw the
line?
MfG Kai
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-REQUEST@lists.debian.org . Trouble? e-mail to Bruce@Pixar.com
Reply to: