From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 22771 invoked by alias); 28 Feb 2004 02:38:36 -0000 Mailing-List: contact gcc-help-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-help-owner@gcc.gnu.org Received: (qmail 22757 invoked from network); 28 Feb 2004 02:38:35 -0000 Received: from unknown (HELO mta6.wss.scd.yahoo.com) (66.218.85.37) by sources.redhat.com with SMTP; 28 Feb 2004 02:38:35 -0000 Received: from specifixinc.com (24.7.123.142) by mta6.wss.scd.yahoo.com (7.0.016) (authenticated as jim@tuliptree.org) id 4029024800A4BB69; Fri, 27 Feb 2004 18:38:33 -0800 Message-ID: <403FFF2D.9090703@specifixinc.com> Date: Sat, 28 Feb 2004 16:06:00 -0000 From: Jim Wilson User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.2) Gecko/20030716 MIME-Version: 1.0 To: aaronchiles@cox.net CC: gcc-help@gcc.gnu.org, gcc@gcc.gnu.org Subject: Re: Building a Fortran Cross Compiler with Gcc-2.95.3 References: <20040227153358.DBQD12901.lakemtao07.cox.net@smtp.central.cox.net> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 8bit X-SW-Source: 2004-02/txt/msg00318.txt.bz2 aaronchiles@cox.net wrote: > After its configures I edit /opt/src/xcomp/binutils-2.14/gas/as.c adding line 38 > 38 #define BFD_VERSION_STRING "2.14" This should not be necessary. BFD_VERSION_STRING comes from bfd/bfdver.h. Why did you add this? >(I chose not to create a build directory because I get the same results and felt I might as well leave it all in the same directory) In general, it really is better to create the separate build directory, even if configuring in the source directory seems to work. You might run into differences later. > CFLAGS = -g 02 -DCPU=PPC604 This is needed only if the powerpc-vxworks target in gcc-2.95 is broken, so this is a bad sign. Adding this to CFLAGS_FOR_TARGET is better than adding it to CFLAGS. You only need the CPU macro defined when compiling stuff for the vxworks target. gcc-2.95 incidentally is pretty old by our standardards. The vxworks stuff was poorly maintained for a long time, so you might have problems with it. The vxworks support has recently been rewritten, and the ppc-vxworks support on the gcc-3.4 branch looks like it might work. I don't think anything before that will work properly without help. > Then I type ‘make’ and it begins to build. After its stop I then type ‘make install’ and it finishes but I believed stopped in the middle. We shouldn't be trying to build anything at make install time. Maybe the initial "make" failed and you didn't notice? > checking how to run C preprocessor… /opt/src/gcc-2.95.3/gcc/xgcc –B/usr/include/sys/ > -B/opt/src/xcomp/gcc-2.95.3/gcc/ -B/opt/src/xcomp/target/powerpc-wrs-vxworks/bin/ -E > checking for stdio.h… no > configure: error: Can’t find stdio.h. > You mus have a usable C system for the target already installed, You should have a usable C compiler at this point. This could be a general vxworks problem, i.e. the vxworks configure support is broken. If you look in the config.log file, you can see the commands that configure ran. It might be trying to link an executable, but you can't do that for a vxworks target, you can only generate partially linked objects. If so, then this is likely fixed in current sources. > I ‘cd’ to /gcc/ and try to ‘make install’ there. Easier is to do "make install-gcc" from the top level. You can also do "make all-gcc" from the top level. > make[2]: Entering directory ‘/opt/src/xcomp/gcc-2.95.3/gcc’ > ./xgcc –B/ opt/src/xcomp/target/powerpc-wrs-vxworks/bin/ -B./ -I/ opt/src/xcomp/target/powerpc-wrs-vxworks/include –DCROSS_COMPILE –DHAIFA –DSVR4 –g –I./include –I. –I. –I./config –I./../include –mrelocatable-lib –mno-eabi –mstrict-align –c ecrti.S –o tecrit.o > make[2]: Leaving directory ‘/opt/src/xcomp/gcc-2.95.3/gcc’ > mv: cannot access tecrti.o There must have been some kind of build error that you didn't report. make install can't find tecrti.o probably because it didn't build. -- Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com