From owner-freebsd-current  Wed Aug  6 23:36:33 1997
Return-Path: <owner-freebsd-current>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id XAA22381
          for current-outgoing; Wed, 6 Aug 1997 23:36:33 -0700 (PDT)
Received: from shale.csir.co.za (shale.csir.co.za [146.64.46.5])
          by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id XAA22376
          for <current@FreeBSD.ORG>; Wed, 6 Aug 1997 23:36:28 -0700 (PDT)
Received: (from reg@localhost)
	by shale.csir.co.za (8.8.6/8.8.6) id WAA10903;
	Wed, 6 Aug 1997 22:54:41 +0200 (SAT)
Message-ID: <19970806225439.18362@shale.csir.co.za>
Date: Wed, 6 Aug 1997 22:54:39 +0200
From: Jeremy Lea <reg@shale.csir.co.za>
To: Satoshi Asami <asami@cs.berkeley.edu>
Cc: current@FreeBSD.ORG
Subject: Re: More bogons in Makefiles...
References: <19894.870331035@time.cdrom.com> <199707311007.DAA14325@silvia.HIP.Berkeley.EDU>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.76
In-Reply-To: <199707311007.DAA14325@silvia.HIP.Berkeley.EDU>; from Satoshi Asami on Thu, Jul 31, 1997 at 03:07:23AM -0700
Sender: owner-freebsd-current@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

Hi,

Finally got around to this...

On Thu, Jul 31, 1997 at 03:07:23AM -0700, Satoshi Asami wrote:
>  * Anyway, that's all for now. I'm still trying to understand make, *.mk and
>  * the Makefiles.
> 
> You know, if you really want to do some work here, try calling the
> source tree something else than "/usr/src" and try to build world from 
> there.  You will be surprised to see just how many places that string
> is hard-coded in.

Only two places broke a 'make world' although it is hard coded in a number
of source files and man pages.

*** ./lkm/atapi/Makefile.orig	Mon Aug  4 22:20:42 1997
--- ./lkm/atapi/Makefile	Wed Aug  6 16:31:08 1997
***************
*** 29,35 ****
  		echo "#define ATAPI 1"> opt_atapi.h
  
  .c.o:
! 		-@$(LN) /sys/i386/include machine
  		$(CC) $(CFLAGS) -c $<
  		-@rm -f machine
  
--- 29,35 ----
  		echo "#define ATAPI 1"> opt_atapi.h
  
  .c.o:
! 		-@$(LN) ${.CURDIR}/../../sys/i386/include machine
  		$(CC) $(CFLAGS) -c $<
  		-@rm -f machine
*** ./lkm/wcd/Makefile.orig	Mon Aug  4 23:02:36 1997
--- ./lkm/wcd/Makefile	Wed Aug  6 16:31:09 1997
***************
*** 25,31 ****
  		echo "#define ATAPI 1"> opt_atapi.h
  
  .c.o:
! 		-@$(LN) /sys/i386/include machine
  		$(CC) $(CFLAGS) -c $<
  		-@rm -f machine
  
--- 25,31 ----
  		echo "#define ATAPI 1"> opt_atapi.h
  
  .c.o:
! 		-@$(LN) ${.CURDIR}/../../sys/i386/include machine
  		$(CC) $(CFLAGS) -c $<
  		-@rm -f machine
  
Both of these were because 'make distrib-dirs' links /sys -> usr/src/sys
instead of to ${.CURDIR}../sys (etc/Makefile line 129). I'm not sure if the
correct behaviour is to always link to /usr/src, because this might be the
"real" src tree for the machine, while the current src tree is meant for
other machines... either way, I think /sys shouldn't be used at all.

 -Jeremy

-- 
.sig.gz