From mboxrd@z Thu Jan  1 00:00:00 1970
From: Daniel De Graaf <dgdegra@tycho.nsa.gov>
Subject: Re: [PATCH 14/14] stubdom/Makefile: Fix gmp extract rule
Date: Mon, 10 Dec 2012 16:28:37 -0500
Message-ID: <50C65405.4070708@tycho.nsa.gov>
References: <1355169347-25917-1-git-send-email-dgdegra@tycho.nsa.gov>
	<1355169347-25917-15-git-send-email-dgdegra@tycho.nsa.gov>
	<20121210212403.GB5833@type.youpi.perso.aquilenet.fr>
Mime-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Return-path: <xen-devel-bounces@lists.xen.org>
In-Reply-To: <20121210212403.GB5833@type.youpi.perso.aquilenet.fr>
List-Unsubscribe: <http://lists.xen.org/cgi-bin/mailman/options/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xen.org>
List-Help: <mailto:xen-devel-request@lists.xen.org?subject=help>
List-Subscribe: <http://lists.xen.org/cgi-bin/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=subscribe>
Sender: xen-devel-bounces@lists.xen.org
Errors-To: xen-devel-bounces@lists.xen.org
To: Samuel Thibault <samuel.thibault@ens-lyon.org>, matthew.fioravante@jhuapl.edu, xen-devel@lists.xen.org
List-Id: xen-devel@lists.xenproject.org

On 12/10/2012 04:24 PM, Samuel Thibault wrote:
> Daniel De Graaf, le Mon 10 Dec 2012 14:55:47 -0500, a =E9crit :
>>  gmp-$(XEN_TARGET_ARCH): gmp-$(GMP_VERSION).tar.bz2 $(NEWLIB_STAMPFILE)
>>  	tar xjf $<
>> +	rm $@ -rf || :
> =

> I'm realizing...
> =

>>  	mv gmp-$(GMP_VERSION) $@
>>  	#patch -d $@ -p0 < gmp.patch
>>  	cd $@; CPPFLAGS=3D"-isystem $(CROSS_PREFIX)/$(GNU_TARGET_ARCH)-xen-elf=
/include $(TARGET_CPPFLAGS)" CFLAGS=3D"$(TARGET_CFLAGS)" CC=3D$(CC) $(GMPEX=
T) ./configure --disable-shared --enable-static --disable-fft --without-rea=
dline --prefix=3D$(CROSS_PREFIX)/$(GNU_TARGET_ARCH)-xen-elf
> =

> Which is configure invoked here?  In the other cases, we have used a
> stampfile, to separate unpacking and configure stages, which avoids the
> issue.  Is it because libgmp does not support build out of tree?  If so,
> then I'm fine with the rm change, although -rm $@ -rf instead of || :
> would be more readable.
> =

> Samuel
> =


I didn't look too closely at the reasons why the configure is done here; the
rm is needed regardless of where the configure is done, assuming the tarball
could be touched while the directory exists.

I chose "|| :" instead of -rm because it silences make's output on the fail=
ure,
which will be normal the first time it is run. If -rm is preferred, that's
trivial to fix.

-- =

Daniel De Graaf
National Security Agency