Re: [ublas] [Lapack Bindings] Strange linker result (i.e. not an unresolved symbol issue)

From: craven (hauke.heibel_at_[hidden])
Date: 2008-08-25 09:25:44

Hi again,

Well, I am not really worried since it does not effect the functionality.
It's just a matter of curiosity and I like clean output directories :).

Regarding the demangling you can use 'dumpbin /EXPORTS <libraryname>' to
extract the symbols and 'undname <mangled name>' for unmangling.

In the meantime I have just found a related thread here:
http://www.tech-archive.net/Archive/VC/microsoft.public.vc.language/2007-02/msg00211.html

I digged a bit deeper into the code and found the 'new (std::nothrow)'
directive that is causing this behavior in 'array_impl.hpp'. Replacing the
code working with 'new (std::nothrow)' by

try {
  sz = n;
  stg = new T[n];
} catch (...) {
  sz = 0;
}

resolved the issue for me. The question is whether that is an acceptable
patch (see  http://www.nabble.com/file/p19143843/array_impl.patch
array_impl.patch ) since I read that on some embedded systems exceptions are
not supported.

Cheers,
Hauke

Thomas Klimpel wrote:
> 
> Hi Hauke,
> 
> I can reproduce this behavior. However, I don't see why this behavior
> should worry me. The test.exp file seems to contain only
> "4_Init_locks_at_std@@QAEAAV01_at_ABV01@@Z" (sorry, I don't know how to
> demangle symbols on Windows) and the test.lib is probably the
> corresponding inline file.
> 
> Regards,
> Thomas
> 

-- 
View this message in context: http://www.nabble.com/-Lapack-Bindings--Strange-linker-result-%28i.e.-not-an-unresolved-symbol-issue%29-tp19139328p19143843.html
Sent from the Boost - uBLAS mailing list archive at Nabble.com.