Subject: Re: bmake doesn't set MAKE correctly?
To: None <firstname.lastname@example.org>
From: Jan Schaumann <email@example.com>
Date: 03/11/2003 18:02:13
"Simon J. Gerraty" <firstname.lastname@example.org> wrote:
> >FWIW, it seems that this happens for packages that USE_GMAKE. So maybe
> >MAKE_PROGRAM and MAKE conflict?
> Ah, that's a good area to look at - generally you need to take care
> to not propagate $MAKE from bmake to gmake environments - else
> bad things can happen. So your problem may well relate to this.
> Though why you only see it on certain platforms is "interesting".
Well, it seems that autoconf/automake also plays a role in this.
When I run 'bmake' in mail/mutt, it configures everything properly, and
then starts to build it. It follows the first dependencies of the
Makefile (namelely the one that says to create Makefile from Makefile.in
I inserted some debugging statement and found:
===> Building for mutt-1.4nb3
/sbin/sed -e "s|@PREFIX@|/usr/pkg|g" -e
./gen_defs ./OPS ./OPS.PGP > keymap_defs.h
./patchlist.sh < ./PATCHES > patchlist.c
echo "MAKE= BUILDLINK_PREFIX.gettext=/usr/pkg
1220 LOWER_OPSYS=irix6 OPSYS=IRIX OS_VERSION=6.5
OWER_OPSYS=irix6 OPSYS=IRIX OS_VERSION=6.5 PKGTOOLS_VERSION=20030202
cd . \
&& CONFIG_FILES=Makefile CONFIG_HEADERS= /bin/ksh ./config.status
So, we find that MAKE is set to what looks like a partial MAKE_ENV
rather than to a single program.
I'll debug this further some time this week, but if anybody else has
insights they'd like to share... :)
It's psychosomatic. You need a lobotomy. I'll get a saw.