It is not a straightforward backport. <atomic> has changed a lot in gcc-4.7. is_lock_free() body ...
12 years, 6 months ago
(2012-07-19 22:48:14 UTC)
#3
It is not a straightforward backport.
<atomic> has changed a lot in gcc-4.7. is_lock_free() body is entirely
different between gcc-4.6 and r183875. In gcc-4.6, is_lock_free()
simply returns false or true. Notice that gcc-4.6 defines two
namesapce __atomic0, __atomic2 in separate files (atomic_0.h,
atomic_2.h), which disappear in gcc-4.7.
Even though, the idea behind the bug is the same. The new unit tests
will fail without this patch.
Thanks,
Jing
Even though, the bug is the same. the new unit tests will
On Thu, Jul 19, 2012 at 3:29 PM, <dougkwan@google.com> wrote:
> This seems to be different from r183875. Are the parts chaing
> is_look_free() in r183875 necessary? If not why?
>
> http://codereview.appspot.com/6428056/
Issue 6428056: [google-4.6]Backport r183875 to fix incorrect increment/decrement of atomic pointers
Created 12 years, 6 months ago by jingyu
Modified 12 years, 6 months ago
Reviewers: Jeffrey Yasskin (google), dougkwan
Base URL: svn+ssh://gcc.gnu.org/svn/gcc/branches/google/gcc-4_6/libstdc++-v3/
Comments: 0