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 diff -urN uClibc-dist/libc/sysdeps/linux/common/bits/uClibc_touplow.h uClibc/libc/sysdeps/linux/common/bits/uClibc_touplow.h
2 --- uClibc-dist/libc/sysdeps/linux/common/bits/uClibc_touplow.h Fri Jun 3 13:30:25 2005
3 +++ uClibc/libc/sysdeps/linux/common/bits/uClibc_touplow.h Fri Jun 3 13:30:39 2005
6 /* glibc uses the equivalent of - typedef __int32_t __ctype_touplow_t; */
8 -typedef __uint16_t __ctype_mask_t;
9 +typedef __uint32_t __ctype_mask_t;
11 #ifdef __UCLIBC_HAS_CTYPE_SIGNED__
13 -typedef __int16_t __ctype_touplow_t;
14 +typedef __int32_t __ctype_touplow_t;
15 #define __UCLIBC_CTYPE_B_TBL_OFFSET 128
16 #define __UCLIBC_CTYPE_TO_TBL_OFFSET 128