From owner-freebsd-ports  Tue Jun  3 17:18:21 1997
Return-Path: <owner-ports>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id RAA03726
          for ports-outgoing; Tue, 3 Jun 1997 17:18:21 -0700 (PDT)
Received: from silvia.HIP.Berkeley.EDU (ala-ca32-09.ix.netcom.com [199.35.209.73])
          by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id RAA03719
          for <ports@FreeBSD.ORG>; Tue, 3 Jun 1997 17:18:19 -0700 (PDT)
Received: (from asami@localhost) by silvia.HIP.Berkeley.EDU (8.8.5/8.6.9) id RAA03524; Tue, 3 Jun 1997 17:18:06 -0700 (PDT)
Date: Tue, 3 Jun 1997 17:18:06 -0700 (PDT)
Message-Id: <199706040018.RAA03524@silvia.HIP.Berkeley.EDU>
To: bde@zeta.org.au
CC: dmaddox@scsn.net, ports@FreeBSD.ORG
In-reply-to: <199706032323.JAA02401@godzilla.zeta.org.au> (message from Bruce Evans on Wed, 4 Jun 1997 09:23:39 +1000)
Subject: Re: Ports 'make install' and mtree
From: asami@cs.berkeley.edu (Satoshi Asami)
Sender: owner-ports@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

 * This is caused by:
 * - /usr/local is owned by bin:bin    according to /etc/mtree/BSD.usr.dist
 * - "."        is owned by root:wheel according to /etc/mtree/BSD.local.dist
 * - `make hierarchy' goes through the mtree files in a strange (alphabetical)
 *   order, so BSD.local.dist is done before BSD.usr.dist.

I didn't know that `make hierarchy' uses BSD.local.dist....

 * It's reasonable for mount points to have different permissions than the
 * root directory of mounted file systems, but the difference is probably
 * backwards here.

Which way should I fix it?  (I was about to ask you, Bruce. :)

 * `make install' runs mtree on /etc/mtree/BSD.local.dist and doesn't run
 * it on BSD.usr.dist.

This is by design.  Unless you can run it on BSD.usr.dist but only
for the "local" directory, I don't think we would want to add that.

 * This is better than pkg_add, which runs mtree on stale copies of
 * BSD.local.dist in packages, so that you can't fix this problem or
 * others in your local copy of BSD.local.dist.

This is again, by design.  The problem is meant to be fixed at the
package-creation machine, not in individual users' machines.

Satoshi