Re: traceroute /usr/bin -> /usr/sbin
Hello,
On Thu, Jun 21, 2001 at 02:43:34PM -0400, Brian Russo wrote:
> since it sparked such.. "debate", FYI:
> i wontfix'd it and will close it eventually,
>
> rationale: consistency
> .. with the rest of the world, not some document.
I'm sorry, but I'm not sure that is an appropriate response. As I have
been pointing out, this is not simply a matter of consistency/adherence with
a document. Debian Policy requires that we follow the FHS (in compatibility
mode). The FHS has clearly required that traceroute (for example) not exist
in /usr/sbin (although the use of wrappers or symlinks to preserve
compatibility is not prohibited since we are working in compatibility mode).
Is it appropriate to "wontfix" a bug that points out that a package is
violating Debian Policy? (ok, it's not explicitly doing that, but that was
its effect.)
> freebsd: /usr/sbin/traceroute
> osf1-4: /usr/sbin/traceroute
> netbsd: /usr/sbin/traceroute
> joe's unix: /usr/sbin/traceroute
> debian/gnu: /usr/sbin/traceroute
How many of these OSses have publicly promised to follow the FHS[1]?
At the moment, I think that we only have two realistic choices for the
proper resolution of this discrepancy (of Debian Policy and practice):
(1) Change the package(s) in question, probably by adding symlinks or
wrappers and eventually deprecating the old location of that is
feasible. This also has the advantage of slaying the "traceroute should
move" thread that keeps returning.
(2) Change Debian Policy to state that we "should" follow the FHS in
compatibility mode (note that there is already a "should" for compliance
mode).
Unfortunately, neither of these seem to be realistic options since we are
heading for a release and Policy is even mostly-frozen[2], I don't think we
will be able to sneak in a revision. If anyone has any other ideas how this
can be fixed, please do speak up (on list or in personal mail to me is
cool), otherwise I suppose we will have to release woody with a policy that
lies about our FHS compliance. Note that potato's debian-policy currently
says that we "must" be compliant as well, so this is not really that new,
it's just the first time I've noticed it, and I appear to be one of the few
people concerned about it (who have said anything that I've seen on the
subject).
As always, alternate interpretations are welcomed.
Rene Weber
References:
[1] <http://www.debian.org/doc/debian-policy/ch-opersys.html> as modified by
<http://bugs.debian.org/98291>
[2] <http://lists.debian.org/debian-policy-0106/msg00080.html>
--
+--- (Rene Weber is <rene_autoreply@elvenlord.com>) ---+
| "Man will occasionally stumble over the truth, but most of the time he |
| will pick himself up and continue on." -- Winston Churchill |
+--- E-Mail Policy & web page: <http://satori.home.dhs.org/~rweber/> ---+
Reply to: