On Apr 09 2019, Carlos O'Donell <codonell@redhat.com> wrote:
https://kojipkgs.fedoraproject.org//work/tasks/2880/34082880/build.log
BUILDSTDERR: In file included from ../sysdeps/aarch64/sysdep.h:22,
BUILDSTDERR: from ../sysdeps/unix/sysv/linux/aarch64/sysdep.h:26,
BUILDSTDERR: from ../sysdeps/aarch64/nptl/tls.h:37,
BUILDSTDERR: from ../include/errno.h:25,
BUILDSTDERR: from ../sysdeps/unix/sysv/linux/netlink_assert_response.c:19:
BUILDSTDERR: ../sysdeps/generic/sysdep.h:81: error: "CFI_RESTORE" redefined [-Werror]
BUILDSTDERR: 81 | # define CFI_RESTORE(reg) \
BUILDSTDERR: |
BUILDSTDERR: In file included from ../sysdeps/unix/sysdep.h:18,
BUILDSTDERR: from ../sysdeps/unix/sysv/linux/aarch64/sysdep.h:25,
BUILDSTDERR: from ../sysdeps/aarch64/nptl/tls.h:37,
BUILDSTDERR: from ../include/errno.h:25,
BUILDSTDERR: from ../sysdeps/unix/sysv/linux/netlink_assert_response.c:19:
BUILDSTDERR: ../sysdeps/generic/sysdep.h:81: note: this is the location of the previous definition
BUILDSTDERR: 81 | # define CFI_RESTORE(reg) \
BUILDSTDERR: |
BUILDSTDERR: cc1: all warnings being treated as errors
This just came up in Fedora Rawhide.
I don't see that failure, and I don't understand why the compiler
complains. The definitions should be 100% identical. Why doesn't the
compiler complain about the other macro redefinitions?