[LLVMdev] linker warnings in Linking CXX executable Debug/AsanTest

[Alexey Samsonov]

Hi, Jack!

On Fri, Nov 2, 2012 at 6:18 PM, Jack Howarth <howarth at bromo.med.uc.edu>wrote:

> Nick,
>    Have you noticed that llvm/clang svn produces the following linker
> warnings on
> 'make check-all'?
>
> Linking CXX executable Debug/AsanTest
> ld: warning: direct access in
> llvm::convertible_fwd_ostream::~convertible_fwd_ostream() to global weak
> symbol vtable for llvm::convertible_fwd_ostream means the weak symbol
> cannot be overridden at runtime. This was likely caused by different
> translation units being compiled with different visibility settings.
> ld: warning: direct access in
> llvm::convertible_fwd_ostream::~convertible_fwd_ostream() to global weak
> symbol vtable for llvm::convertible_fwd_ostream means the weak symbol
> cannot be overridden at runtime. This was likely caused by different
> translation units being compiled with different visibility settings.
> ld: warning: direct access in non-virtual thunk to
> llvm::convertible_fwd_ostream::~convertible_fwd_ostream() to global weak
> symbol vtable for llvm::convertible_fwd_ostream means the weak symbol
> cannot be overridden at runtime. This was likely caused by different
> translation units being compiled with different visibility settings.
> ld: warning: direct access in non-virtual thunk to
> llvm::convertible_fwd_ostream::~convertible_fwd_ostream() to global weak
> symbol vtable for llvm::convertible_fwd_ostream means the weak symbol
> cannot be overridden at runtime. This was likely caused by different
> translation units being compiled with different visibility settings.
> ld: warning: direct access in
> llvm::convertible_fwd_ostream::~convertible_fwd_ostream() to global weak
> symbol vtable for llvm::convertible_fwd_ostream means the weak symbol
> cannot be overridden at runtime. This was likely caused by different
> translation units being compiled with different visibility settings.
> ld: warning: direct access in
> llvm::convertible_fwd_ostream::~convertible_fwd_ostream() to global weak
> symbol vtable for llvm::convertible_fwd_ostream means the weak symbol
> cannot be overridden at runtime. This was likely caused by different
> translation units being compiled with different visibility settings.
> ld: warning: direct access in non-virtual thunk to
> llvm::convertible_fwd_ostream::~convertible_fwd_ostream() to global weak
> symbol vtable for llvm::convertible_fwd_ostream means the weak symbol
> cannot be overridden at runtime. This was likely caused by different
> translation units being compiled with different visibility settings.
> ld: warning: direct access in non-virtual thunk to
> llvm::convertible_fwd_ostream::~convertible_fwd_ostream() to global weak
> symbol vtable for llvm::convertible_fwd_ostream means the weak symbol
> cannot be overridden at runtime. This was likely caused by different
> translation units being compiled with different visibility settings.
>

I'll take a look at this. However, tests below fail for a different reason
(they don't use Debug/AsanTest at all).
How do you configure a CMake build tree? Can you somehow get (or run
manually) the script which is used
for running log-path_test.cc and other failing test cases?


>
> I am seeing some AddressSanitizer failures like...
>
> FAIL: AddressSanitizer :: log-path_test.cc (13 of 12542)
> ******************** TEST 'AddressSanitizer :: log-path_test.cc' FAILED
> ********************
>
> FAIL: AddressSanitizer :: strncpy-overflow.cc (27 of 12542)
> ******************** TEST 'AddressSanitizer :: strncpy-overflow.cc' FAILED
> ********************
>
> FAIL: AddressSanitizer :: use-after-free.cc (34 of 12542)
> ******************** TEST 'AddressSanitizer :: use-after-free.cc' FAILED
> ********************
>
> etc, and wonder if they could be related to that those linkage warnings.
> If so, I'll open a bug report.
> Thanks in advance.
>          Jack
> ps This is using a debug build on x86_64-apple-darwin12 with Xcode 4.5.2's
> linker.
>
>
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>



-- 
Alexey Samsonov, MSK
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20121102/44ac2980/attachment.html>