From owner-freebsd-commit  Thu Oct 19 21:07:33 1995
Return-Path: owner-commit
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id VAA14147
          for freebsd-commit-outgoing; Thu, 19 Oct 1995 21:07:33 -0700
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id VAA14131
          for cvs-all-outgoing; Thu, 19 Oct 1995 21:07:28 -0700
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id VAA14114
          for cvs-CVSROOT-outgoing; Thu, 19 Oct 1995 21:07:24 -0700
Received: from veda.is (root@veda.is [193.4.230.1])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id VAA14096
          ; Thu, 19 Oct 1995 21:07:04 -0700
Received: (from adam@localhost) by veda.is (8.6.12/8.6.12) id EAA04706; Fri, 20 Oct 1995 04:06:41 GMT
From: Adam David <adam@veda.is>
Message-Id: <199510200406.EAA04706@veda.is>
Subject: Re: cvs commit: CVSROOT avail
To: asami@cs.berkeley.edu (Satoshi Asami)
Date: Fri, 20 Oct 1995 04:06:34 +0000 (GMT)
Cc: CVS-commiters@freefall.freebsd.org, cvs-CVSROOT@freefall.freebsd.org
In-Reply-To: <199510200325.UAA02998@silvia.HIP.Berkeley.EDU> from "Satoshi Asami" at Oct 19, 95 08:25:17 pm
X-Mailer: ELM [version 2.4 PL24 ME8a]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1865      
Sender: owner-commit@FreeBSD.org
Precedence: bulk

> By the way, why is the wn port NO_PACKAGE?  Also, the @exec statement
> in pkg/PLIST, what is it supposed to do?  Don't you need a
> corresponding @unexec statement to clean up the files during
> pkg_delete?
> 
> Satoshi
> 

The default mailto:webmaster@your.site URL is compiled into the binaries with
local site info. It would be possible to make a generic package by leaving it
as a bogus address, but this requires the directories of the data heirarchy to
specify a maintainer (which should be optional). I've no idea why the author didn't
use a startup config file for this purpose instead of compiling it in.

Also, it is extra work to cleanly create the package on a system which already
has a data heirarchy installed (probably modified locally). This is less of a
hinderance than the previous reason.

The missing @unexec is probably an oversight on my part. originally I just
named the whole server tree, but decided that it would too easily pull in
locally supplied files in the current situation of possibly trying to build
a package from an installed data heirarchy. Eventually, when the package
can be built from a virgin build without being installed first, it might
make better sense to specify 'wn' as a single line again, instead of both
'wn/index' and 'wn/docs'.

The @exec statement builds the cache files and the root index.html file.

Also, there is the consideration of minimum damage, in the case where someone
might want to delete and reinstall a package. The data heirarchy provided is a
skeleton which is intended as an example to be modified and added to. Obviously,
production servers will specify a separate data heirarchy location, but there
are those who might neglect to do this.

I will give it more thought when I get it to make a package by default. There
is currently a method of doing this, but it's not very pretty :)

Adam