From owner-freebsd-current Tue Jun 9 22:14:57 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id WAA24081 for freebsd-current-outgoing; Tue, 9 Jun 1998 22:14:57 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from cimlogic.com.au (cimlog.lnk.telstra.net [139.130.51.31]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id WAA24032 for ; Tue, 9 Jun 1998 22:14:49 -0700 (PDT) (envelope-from jb@cimlogic.com.au) Received: (from jb@localhost) by cimlogic.com.au (8.8.8/8.8.7) id PAA13544; Wed, 10 Jun 1998 15:18:23 +1000 (EST) (envelope-from jb) From: John Birrell Message-Id: <199806100518.PAA13544@cimlogic.com.au> Subject: Re: Can't "make buildworld" In-Reply-To: <199806100502.PAA21808@godzilla.zeta.org.au> from Bruce Evans at "Jun 10, 98 03:02:26 pm" To: bde@zeta.org.au (Bruce Evans) Date: Wed, 10 Jun 1998 15:18:23 +1000 (EST) Cc: bde@zeta.org.au, jb@cimlogic.com.au, freebsd-current@FreeBSD.ORG, wpub1@triton.net X-Mailer: ELM [version 2.4ME+ PL40 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Bruce Evans wrote: > The top level is supposed to do very little, so that the files in > /usr/share/mk are good enough. The point I'm concerned about is that > when -I is used to force use of src/share/*.mk for building `make', > it may give a conflicting set of .mk files - sys.mk from /usr/share/mk > and everything else source-relative. At least one other bug has crept > in: the source-relative files are not used at all for bootstrapping > mtree. The problem reported was the evaluation of MACHINE_ARCH (in src/Makefile) on i386 which is set in sys.mk because historically FreeBSD's make has not had this compiled in despite the fact that the source supports it. So testing MACHINE_ARCH in src/Makefile without using /usr/src/share/mk/sys.mk results in MACHINE_ARCH being undefined and the build blows up before it gets to build make or anything else. It was my recent addition of the MACHINE_ARCH test in src/Makefile that (a) allowed src/Makefile to build world on alpha; and (b) caused the build world on oldish systems to fail. In this case, I don't think the make ever gets to the point of being able to make 'make' - with or without the -I. I guess the conclusion then is that the top level is doing too much. -- John Birrell - jb@cimlogic.com.au; jb@freebsd.org http://www.cimlogic.com.au/ CIMlogic Pty Ltd, GPO Box 117A, Melbourne Vic 3001, Australia +61 418 353 137 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message