Bart De VOS pointed out that removing absolute paths from the libc linker scripts is plainly wrong.
It dates from dawn ages of the original crosstool code, and is not well explained. At that time, binutils might not understand the sysroot stuff, and it was necessary to remove absolute paths in that case.
/trunk/scripts/build/libc/glibc.sh | 14 2 12 0 ++------------
1 file changed, 2 insertions(+), 12 deletions(-)
1 ML: http://sources.redhat.com/ml/libc-hacker/2005-09/msg00002.html
3 Date: Mon, 05 Sep 2005 21:07:15 +0900 (JST)
4 Message-Id: <20050905.210715.15267870.kkojima@rr.iij4u.or.jp>
5 To: libc-hacker at sources dot redhat dot com
6 Subject: SH: A typo in lowlevellock.S
7 From: Kaz Kojima <kkojima at rr dot iij4u dot or dot jp>
11 The appended patch fixes a typo in a low-level lock function. It
12 set the correct 3rd argument for the futex syscall in loop. Sorry
13 for missing such an embarrassing bug.
19 2005-09-05 Kaz Kojima <kkojima@rr.iij4u.or.jp>
21 * sysdeps/unix/sysv/linux/sh/lowlevellock.S (__lll_mutex_lock_wait):
22 Fix typo in register name.
24 --- glibc.old/nptl/sysdeps/unix/sysv/linux/sh/lowlevellock.S 2004-10-26 04:06:44.000000000 +0900
25 +++ glibc/nptl/sysdeps/unix/sysv/linux/sh/lowlevellock.S 2005-09-05 19:18:25.000000000 +0900
27 -/* Copyright (C) 2003, 2004 Free Software Foundation, Inc.
28 +/* Copyright (C) 2003, 2004, 2005 Free Software Foundation, Inc.
29 This file is part of the GNU C Library.
31 The GNU C Library is free software; you can redistribute it and/or
32 @@ -51,8 +51,8 @@ __lll_mutex_lock_wait:
43 Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>