stdcxx-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Martin Sebor <>
Subject Re: ATOMIC_OPS.cpp on MSVC8.0-x64
Date Mon, 27 Aug 2007 18:52:09 GMT
Farid Zaripov wrote:
>   ATOMIC_OPS.cpp configuration test failed on MSVC8.0-x64
> due to undefined external symbol _InterlockedIncrement.
> Because of this _RWSTD_NO_ATOMIC_OPS macro has
> been #defined in config.h.
>   The InterlockedIncrement() function defined only in Win32 platform
> (located in kernel32.dll). At x64 platform this function is declared,
> but kernel32.dll doesn't exports this function. The MSVC8 injects
> code of the function instead of call (so called intrinsic function).

Does Microsoft document that it's illegal to take the address
of an intrinsic function as the ATOMIC_OPS.cpp test does? If
not, it's a bug in their compiler (we need to work around it,
but we should also file it with them).


View raw message