From owner-freebsd-bugs  Wed Jan 15 05:33:25 1997
Return-Path: <owner-bugs>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.4/8.8.4) id FAA01141
          for bugs-outgoing; Wed, 15 Jan 1997 05:33:25 -0800 (PST)
Received: from squirrel.tgsoft.com (squirrel.tgsoft.com [207.167.64.183])
          by freefall.freebsd.org (8.8.4/8.8.4) with ESMTP id FAA01117
          for <freebsd-bugs@freefall.freebsd.org>; Wed, 15 Jan 1997 05:33:22 -0800 (PST)
Received: (from thompson@localhost) by squirrel.tgsoft.com (8.8.3/8.6.12) id FAA21208; Wed, 15 Jan 1997 05:34:49 -0800 (PST)
Date: Wed, 15 Jan 1997 05:34:49 -0800 (PST)
Message-Id: <199701151334.FAA21208@squirrel.tgsoft.com>
From: mark thompson <thompson@tgsoft.com>
To: jkh@time.cdrom.com
CC: freebsd-bugs@freefall.freebsd.org
In-reply-to: <21415.853307483@time.cdrom.com> (jkh@time.cdrom.com)
Subject: Re: bin/2493: make $DESTDIR work
Sender: owner-bugs@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

   cc: freebsd-bugs@freefall.freebsd.org
   Date: Tue, 14 Jan 1997 21:51:23 -0800
   From: "Jordan K. Hubbard" <jkh@time.cdrom.com>

   >  I believe that whether or not ${DESTDIR} was 'guaranteed' to do what i
   >  want (allow you to do a make world without overwritting your running
   >  system), 90% of the Makefiles are coded so you can do just that. The
   >  changes that i submitted take you to between 95-99% of the way there and
   >  *should not* break anything. 

   I dunno, sounds convincing to me, Joerg. ;-)

   How much testing have you done with this, Mark?  When you say *should
   not* is this from long-standing experience and much comparative
   testing of binaries (run mtree over tree A, then over tree B, compare
   results) or is this more of an engineer's standard "*I hope it won't*"
   :-)

   If it's the former, I certainly don't see a problem.  I agree that
   there's no reason NOT to make DESTDIR a more capable instrument
   if it doesn't break any existing functionality.

						   Jordan

Ah. That is the rub. I have successfully used just what i sent in to build
the system (many times, to get it reliable), and have used *that* to
build itself in my traditional way (chroot to new system and do a
make world). I plan do an A-B comparision, but (for the usual excuses)
haven't as of yet.  
You will recall that when i raised this whole business week or so ago,
you suggested that whatever patch i had might be more useful as an
example than something to actually use. In the event, some of the
changes are "obviously" ok, and, sadly several are not so obvious. 
Furthermore, the games subdirectory still has a problem or two (i know
that there are at least some malloc faux-pas).
I leave it to the judgement of the court. Certainly, it is possible that
i could do more development on this patch if the value is really there.

-mark