From owner-freebsd-current  Sun Sep 22 23:36:38 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id XAA17942
          for current-outgoing; Sun, 22 Sep 1996 23:36:38 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id XAA17823
          for <freebsd-current@freefall.freebsd.org>; Sun, 22 Sep 1996 23:36:24 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.7.6/8.6.9) id QAA07882; Mon, 23 Sep 1996 16:32:10 +1000
Date: Mon, 23 Sep 1996 16:32:10 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199609230632.QAA07882@godzilla.zeta.org.au>
To: ache@nagual.ru, chuckr@glue.umd.edu, steve@freebsd.org
Subject: Re: Make world
Cc: freebsd-current@freefall.freebsd.org
Sender: owner-current@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>>> ===> usr.bin/fpr
>>> Warning: Using /usr/obj/usr/src/usr.bin/fpr as object directory instead of
>>> canonical /usr/obj/usr/src/usr.bin/fpr
>>
>>They means that obj directories exists in each case and you not turn it
>>on directly in make.conf
>
>Even when obj directory that is used has the same name as the canonical
>one? :-)  I haven't seen this.  It can't happen if the names are really
>identical.

I see this now.  Operator precedeence in .elif seems to have been broken
by yesterday's changes.

Try the following Makefile.  The .elif test succeeds.

Bruce

---
ONE=one

foo:
.if !defined(NOTDEF) && ${ONE} != ${ONE}
	@echo "This shouldn't be printed"
.elif !defined(NOTDEF) && ${ONE} != ${ONE}
	@echo "This shouldn't be printed either"
.endif
---