From owner-freebsd-current  Tue Apr  4 11:34:23 1995
Return-Path: current-owner
Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id LAA19277 for current-outgoing; Tue, 4 Apr 1995 11:34:23 -0700
Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id LAA19264 for <current@FreeBSD.org>; Tue, 4 Apr 1995 11:34:16 -0700
Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.8/8.6.6) id KAA07885; Tue, 4 Apr 1995 10:33:05 -0700
From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com>
Message-Id: <199504041733.KAA07885@gndrsh.aac.dev.com>
Subject: Re: cvs commit: src/sys/i386/conf Makefile.i386
To: nate@trout.sri.MT.net (Nate Williams)
Date: Tue, 4 Apr 1995 10:33:05 -0700 (PDT)
Cc: rkw@dataplex.net, current@FreeBSD.org
In-Reply-To: <199504041758.LAA06600@trout.sri.MT.net> from "Nate Williams" at Apr 4, 95 11:58:51 am
X-Mailer: ELM [version 2.4 PL23]
Content-Type: text
Content-Length: 1514      
Sender: current-owner@FreeBSD.org
Precedence: bulk

> 
> > >But, *IF* both the src dist and /usr/include directories exist, the
> > >include files are pulled out of /usr/src/include if they exist and
> > >/usr/include next.  (Which should never happen since /usr/src/include
> > >should always contain all of the necessary include files which aren't in
> > >the /usr/src/sys tree)
> > 
> > I think that this is bogus behavior.
> > 
> > If a file is missing from /usr/src/include, I don't want it to be
> > "magically" found.
> 
> First of all, there are only two include files that are used that aren't
> in /usr/src/sys (for the genassym program and for one source file), so
> it's not a critical problem, and second of all if you don't want to
> worry about /usr/include stuff being picked up, do a 'mv /usr/include
> /usr/include.old' before building your kernel.

The rule for genassym's should be changed such that it always uses
/usr/include/stdio.h, that is the correct file to use, period, as
it is going to be linked and run against /usr/lib/libc.a at the time
the kernel is compiled.  This will be extreamly important for cross
compilation in the future.

...

For now I am willing to live with what Nate did do the impact it has
on our standard user base to do otherwise.  We do need to look at
better ways for fixing this, but as Nate said, until then this is a
reasonable way to solve the problem.

-- 
Rod Grimes                                      rgrimes@gndrsh.aac.dev.com
Accurate Automation Company                   Custom computers for FreeBSD