Return-Path: mao
Received: by postgres.Berkeley.EDU (5.61/1.29)
	id AA03348; Tue, 11 Feb 92 18:16:20 -0800
Message-Id: <9202120216.AA03348@postgres.Berkeley.EDU>
From: mao@postgres.Berkeley.EDU (Mike Olson)
Subject: Re: postgres on SCO Unix (fwd)
To: postgres@postgres.berkeley.edu
Sender: pg_adm@postgres.berkeley.edu
In-Reply-To: Your message of Tue, 11 Feb 92 14:34:25 EST.
             <9202112343.AA00401@postgres.Berkeley.EDU> 
Date: Tue, 11 Feb 92 18:09:12 PST

In message <9202112343.AA00401@postgres.Berkeley.EDU> you write:

> First: I had to create a modified version of GNU Make.  The SysV make
> and GNU Make by default had trouble feeding large command lines to the
> shell, apparently based on a kernel argv buffer limit.  I used pipe or
> something similar.  I still have death of child problems, which are
> ignorable.

the problem in gnu make is that by default, it exports the names and
values of make variables, which in postgres get very large.  i had
to deal with this in our sequent port recently.  in the version of
gnu make i used, you can define a pseudo-target

	.NOEXPORT:

in the config.mk file that will cause make variables not to be exported.
this was the only change i needed to make to gnu make to get it to
eat postgres makefiles successfully.

i realize that this is a little late for you, but in case anyone else
is interested in doing this work, thought i'd send it out.

by the way, if anyone knows anyone at the fsf, you might mention that
this is a serious misfeature of gnu make.

					mike olson
					postgres research group
					uc berkeley
					mao@cs.berkeley.edu