[LLVMbugs] [Bug 12730] libstdc++ std::thread, when compiled with clang, "pure virtual method called"

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=12730

Niall Douglas <s_llvm at nedprod.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
                 CC|                            |s_llvm at nedprod.com
         Resolution|FIXED                       |---

--- Comment #13 from Niall Douglas <s_llvm at nedprod.com> ---
@Michael Spencer: Is there a good reason why your fix adding the
__GCC_HAVE_SYNC_COMPARE_AND_SWAP_[1248] macros in SVN rev 178816 is x86 only?

If you look at http://sourceware.org/ml/libc-ports/2010-11/msg00005.html you'll
see that on ARM the use of atomic builtins only happens if
__GCC_HAVE_SYNC_COMPARE_AND_SWAP_4 is set.

This rather implies the need for every architecture target, not just x86, to
set __GCC_HAVE_SYNC_COMPARE_AND_SWAP_[1248] macros according to whether that
target provides CAS ops of varying sizes.

FYI I just got bit on this bug but with an ARM target, hence me finding this
report.

Niall

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20130715/ed3ec486/attachment.html>