From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-278215-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 27130 invoked by alias); 12 Nov 2010 15:02:54 -0000
Received: (qmail 27116 invoked by uid 22791); 12 Nov 2010 15:02:53 -0000
X-SWARE-Spam-Status: No, hits=-2.5 required=5.0	tests=AWL,BAYES_00,RCVD_IN_DNSWL_LOW
X-Spam-Check-By: sourceware.org
Received: from c60.cesmail.net (HELO c60.cesmail.net) (216.154.195.49)    by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Fri, 12 Nov 2010 15:02:06 +0000
Received: from unknown (HELO epsilon2) ([192.168.1.60])  by c60.cesmail.net with ESMTP; 12 Nov 2010 10:02:04 -0500
Received: from 89.241.153.155 ([89.241.153.155]) by webmail.spamcop.net	(Horde MIME library) with HTTP; Fri, 12 Nov 2010 10:02:04 -0500
Message-ID: <20101112100204.lhw5uyhrpcswogo8-nzlynne@webmail.spamcop.net>
Date: Fri, 12 Nov 2010 15:46:00 -0000
From: Joern Rennecke <amylaar@spamcop.net>
To: Paolo Bonzini <bonzini@gnu.org>
Cc: gcc-patches@gcc.gnu.org, DJ Delorie <dj@redhat.com>, Alexandre Oliva	<aoliva@redhat.com>
Subject: Re: Ping^2: RFA (build): Avoid circular dependency
References: <20100629172139.7sxigvan4g48koss-nzlynne@webmail.spamcop.net>	<20101104090834.yzg2549rks8o84wg-nzlynne@webmail.spamcop.net>	<AANLkTimFkq61sweWhQ+qY2H2Muu6=hkeNFRkfPp0A+wh@mail.gmail.com>	<20101112091310.xfc3pvh9ogw0sskg-nzlynne@webmail.spamcop.net>	<4CDD502D.9060405@gnu.org>
In-Reply-To: <4CDD502D.9060405@gnu.org>
MIME-Version: 1.0
Content-Type: text/plain;	charset=UTF-8;	DelSp="Yes";	format="flowed"
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
User-Agent: Internet Messaging Program (IMP) H3 (4.1.4)
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-patches.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
X-SW-Source: 2010-11/txt/msg01297.txt.bz2

Quoting Paolo Bonzini <bonzini@gnu.org>:

> On 11/12/2010 03:13 PM, Joern Rennecke wrote:
...
>> Are you still planning to do the review in the near future?
>
> Yes.  Your patch has the disadvantage that a normal "make" will not
> do the check anymore, because you removed the dependency from all to
> s-tm-texi (which previously went all -> info -> gccint.info ->
> tm.texi -> s-tm-texi).

We still have a gccint.info -> new-tm.texi -> s-tm-texi link.

> What about this instead, which breaks the circular dependency by
> making tm.texi phony.  This way "$(MAKE) tm.texi" is effectively
> a synonym of "$(MAKE s-tm-texi)".
>
> --- Makefile.in	(revision 166517)
> +++ Makefile.in	(working copy)
> @@ -3688,7 +3688,10 @@ s-constrs-h: $(MD_DEPS) build/genpreds$(
>  	$(STAMP) s-constrs-h
>
>  target-hooks-def.h: s-target-hooks-def-h; @true
> -tm.texi: s-tm-texi; @true
> +
> +all: tm.texi
> +tm.texi: force
> +	$(MAKE) s-tm-texi

This looks unsafe in a parallel make context.
make could decide to build tm.texi and s-tm-texi simultaneously.
Then in order to build tm.texi, it will invoke another make that
builds s-tm-texi, too.