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 -durN gcc-4.2.1.orig/gcc/stmt.c gcc-4.2.1/gcc/stmt.c
2 --- gcc-4.2.1.orig/gcc/stmt.c 2007-07-19 05:25:32.000000000 +0200
3 +++ gcc-4.2.1/gcc/stmt.c 2007-08-03 20:37:19.000000000 +0200
6 = (TREE_CODE (orig_type) != ENUMERAL_TYPE
7 && estimate_case_costs (case_list));
8 - balance_case_nodes (&case_list, NULL);
9 + /* When optimizing for size, we want a straight list to avoid
10 + jumps as much as possible. This basically creates an if-else
13 + balance_case_nodes (&case_list, NULL);
14 emit_case_nodes (index, case_list, default_label, index_type);
15 emit_jump (default_label);
19 if (!node_has_low_bound (node, index_type))
21 + if (!optimize_size) /* don't jl to the .default_label. */
22 emit_cmp_and_jump_insns (index,