From mboxrd@z Thu Jan  1 00:00:00 1970
X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/475
Path: news.gmane.org!not-for-mail
From: Rich Felker <dalias@aerifal.cx>
Newsgroups: gmane.linux.lib.musl.general
Subject: New gcc wrapper, take 3
Date: Wed, 21 Sep 2011 16:41:04 -0400
Message-ID: <20110921204104.GH132@brightrain.aerifal.cx>
Reply-To: musl@lists.openwall.com
NNTP-Posting-Host: lo.gmane.org
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="kvUQC+jR9YzypDnK"
X-Trace: dough.gmane.org 1316638000 21738 80.91.229.12 (21 Sep 2011 20:46:40 GMT)
X-Complaints-To: usenet@dough.gmane.org
NNTP-Posting-Date: Wed, 21 Sep 2011 20:46:40 +0000 (UTC)
To: musl@lists.openwall.com
Original-X-From: musl-return-476-gllmg-musl=m.gmane.org@lists.openwall.com Wed Sep 21 22:46:36 2011
Return-path: <musl-return-476-gllmg-musl=m.gmane.org@lists.openwall.com>
Envelope-to: gllmg-musl@lo.gmane.org
Original-Received: from mother.openwall.net ([195.42.179.200])
	by lo.gmane.org with smtp (Exim 4.69)
	(envelope-from <musl-return-476-gllmg-musl=m.gmane.org@lists.openwall.com>)
	id 1R6Tgi-0002kT-NI
	for gllmg-musl@lo.gmane.org; Wed, 21 Sep 2011 22:46:36 +0200
Original-Received: (qmail 21999 invoked by uid 550); 21 Sep 2011 20:46:36 -0000
Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm
Precedence: bulk
List-Post: <mailto:musl@lists.openwall.com>
List-Help: <mailto:musl-help@lists.openwall.com>
List-Unsubscribe: <mailto:musl-unsubscribe@lists.openwall.com>
List-Subscribe: <mailto:musl-subscribe@lists.openwall.com>
Original-Received: (qmail 21991 invoked from network); 21 Sep 2011 20:46:36 -0000
Content-Disposition: inline
User-Agent: Mutt/1.5.21 (2010-09-15)
Xref: news.gmane.org gmane.linux.lib.musl.general:475
Archived-At: <http://permalink.gmane.org/gmane.linux.lib.musl.general/475>


--kvUQC+jR9YzypDnK
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

Now updated to use -specs instead of extra wrapping layers with -B in
order to inhibit the default library paths. This version should be
faster, cleaner, and perhaps even more correct than the previous
attempts. I believe some of the other options could also be integrated
into the specs file passed in, and this might make things even cleaner
still.

Please test and let me know if there are problems.

Rich

--kvUQC+jR9YzypDnK
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename=my-musl-gcc

#!/bin/sh

gcc=gcc
libc_prefix="/usr/local/musl"
ldso_pathname="/lib/ld-musl-i386.so.1"
libc_lib=$libc_prefix/lib
libc_inc=$libc_prefix/include
libc_crt="$libc_lib/crt1.o"
libc_start="$libc_lib/crti.o"
libc_end="$libc_lib/crtn.o"

gcc_inc=$libc_inc
libgcc="$("$gcc" -print-file-name=libgcc.a)"
libgcc=${libgcc%/libgcc.a}

gccver=${libgcc##*/}
gcctarget=${libgcc%/*}
gcctarget=${gcctarget##*/}

case "$gccver" in
2.*|3.*|4.[01]*) ;;
*) nosp=-fno-stack-protector ;;
esac

[ "x$1" = "x-V" ] && { printf "%s: -V not supported\n" "$0" ; exit 1 ; }

for i ; do
case "$skip$i" in
-I|-L) skip=--- ; continue ;;
-[cSE]|-M*) nolink=1 ;;
-*) ;;
*) havefile=1 ;;
esac
skip=
done

[ "$havefile" ] || nolink=1

[ "$nolink" ] || {
tmp_specs=$HOME/.specs.tmp.$$
printf '*link_libgcc:\n\n\n' > "$tmp_specs" || exit 1
set -- -specs="$tmp_specs" "$libc_start" "$libc_crt" "$@" "$libc_end" \
  -Wl,--start-group -lc -lgcc -lgcc_eh -Wl,--end-group \
  -Wl,-dynamic-linker,"$ldso_pathname" -Wl,-nostdlib
}

set -- -std=gnu99 -nostdinc -nostdlib $nosp \
  -isystem "$libc_inc" -isystem "$gcc_inc" "$@" \
  -L"$libc_lib" -L"$libgcc"

"$gcc" "$@"

[ "$tmp_specs" ] && rm -f "$tmp_specs"

--kvUQC+jR9YzypDnK--