From owner-freebsd-current  Wed Jun  5 04:30:19 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id EAA16405
          for current-outgoing; Wed, 5 Jun 1996 04:30:19 -0700 (PDT)
Received: from diablo.ppp.de (diablo.ppp.de [193.141.101.34])
          by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id EAA16392
          for <FreeBSD-current@FreeBSD.ORG>; Wed, 5 Jun 1996 04:30:15 -0700 (PDT)
Received: from allegro.lemis.de by diablo.ppp.de with smtp
	(Smail3.1.28.1 #1) id m0uRGjd-000QZQC; Wed, 5 Jun 96 13:26 MET DST
From: grog@lemis.de (Greg Lehey)
Organisation: LEMIS, Schellnhausen 2, 36325 Feldatal, Germany
Phone: +49-6637-919123
Fax:   +49-6637-919122
Received: (grog@localhost) by allegro.lemis.de (8.6.9/8.6.9) 
	id NAA08452 for FreeBSD-current@FreeBSD.ORG; Wed, 5 Jun 1996 13:13:19 +0200
Message-Id: <199606051113.NAA08452@allegro.lemis.de>
Subject: Re: stable breakage
To: randy@zyzzyva.com (Randy Terbush)
Date: Wed, 5 Jun 1996 13:12:50 +0200 (MET DST)
In-Reply-To: <199606012008.PAA01451@sierra.zyzzyva.com> from "Randy Terbush" at Jun 1, 96 03:08:12 pm
X-Mailer: ELM [version 2.4 PL23]
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
Sender: owner-current@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

Randy Terbush writes:
>
>>> rune.c isn't needed anymore as it only contains the functions
>>> setinvalidrune() and setrunelocale(), which are defined in setinvalidrune.c
>>> and setrunelocale.c.
>>>
>>> Remove the reference to runc.c in lib/libc/locale/Makefile.inc to fix this
>>> problem.  (You can nuke rune.c as well.)
>>
>> Huh?  Are you sure?  It's still in the 2.2-current version of this
>> file.  I'd love to have the fix be so simple, but if this is still
>> done in 2.2 then I have to wonder..
>
> Nope! Not quite that simple. It seems that common_setlocale.c also needs
> to be added to the libc/locale/Makefile.inc. I seriously hosed myself
> doing make world when new_categories() was suddenly missing.
>
> A quick comparison of the libc.so.2.2 from 2.1R shows that there may
> be other things missing as well.
>
>> From 2.1R
> -r--r--r--  1 bin  bin  435248 Jun  1 14:25 /usr/lib/libc.so.2.2
> From the current -stable
> -rwxr-x---  1 root  develop  384472 Jun  1 15:01 /usr/src/lib/libc/obj/libc.so.2
>
> Word of warning:
> Unless you are linking everything static, a make world in -stable may
> seriously ruin your day.

Tell me about it.  It's a good thing to have another box NFS mounted.

I'm really making -stable for installation on another system.  I can't
see any good reason to have to install it on the system on which I'm
building in order to achieve this goal.  I've already modified the
libc Makefiles to take header files only from the /usr/src hierarchy;
it should be possible to use different binary and library paths as
well, so that the build environment is completely separate from the
host environment.  I know there's been a lot of talk about this
before, but what are the current arguments against such a solution?

Greg