Hi, On Wed, Apr 23, 2014 at 8:27 PM, Allan McRae <allan@archlinux.org> wrote:

I have pushed the gcc-4.9 toolchain to the [testing] repo. There are a couple of testsuite failures remaining (detailed below). The binutils one I think is a false positive, the gcc one is unimportant. I am still unsure about the glibc issue and will discuss upstream.

One this is out of [testing], I will change the CFLAGS from -fstack-protector to -fstack-protector-strong in our makepkg.conf.

Changes:

gcc-4.9.0-1: - new upstream release - http://gcc.gnu.org/gcc-4.9/changes.html - currently small test failure on x86_64 only (c-c++-common/tsan/thread_leak1.c)

linux-api-headers-3.14.1-1: - upstream update - add upstream patch to fix dual xattr.h inclusion

binutils-2.24-3: - add upstream patch testsuite failures with gcc-4.9 - one failure (LTO 11) remains

glibc-2.19-4: - replace hack fix for dual xattr.h inclusion with upstream one - i686 has minor failures in libm ULPs (patch sent upstream) - i686 has a failure in nptl/tst-cleanupx4.out. No idea on that one yet...

libtool-2.4.2-13: - rebuild

I am trying to use gdc in [testing] and it fails: gdc: fatal error: -fuse-linker-plugin, but liblto_plugin.so not found I ran strace for gdc command and it shows it tries to find liblto in the old version locations. It seems gdc statically compiled against gcc and should be recompiled against 4.9 as well. access("/usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.2/liblto_plugin.so", R_OK) = -1 ENOENT (No such file or directory) access("/usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.2/liblto_plugin.so", R_OK) = -1 ENOENT (No such file or directory) access("/usr/lib/gcc/x86_64-unknown-linux-gnu/liblto_plugin.so", R_OK) = -1 ENOENT (No such file or directory) access("/usr/lib/gcc/x86_64-unknown-linux-gnu/4.8.2/liblto_plugin.so", R_OK) = -1 ENOENT (No such file or directory) access("/usr/lib/gcc/x86_64-unknown-linux-gnu/liblto_plugin.so", R_OK) = -1 ENOENT (No such file or directory)