[cfe-dev] Compiler error including windows.h

[Nico Weber via cfe-dev]

It's a bug in the mingw headers. If you pass -fms-extensions
-fms-extensions, clang will accept that code since Microsoft's headers have
the same bug.

On May 7, 2017 10:05 PM, "Edward Diener via cfe-dev" <cfe-dev at lists.llvm.org>
wrote:

> I build clang from the latest source using mingw-64/gcc-7.1. Using
> mingw-64/gcc-7.1 as the backend, when I compile:
>
> // test_compiler.cpp
>
> #include <windows.h>
> int main()
> {
> return 0;
> }
>
> with the command line:
>
> some_path/clang++" -c -x c++ -O0 -g -fno-inline -Wall -g -march=i686 -m32"
> -o "some_path\test_compiler.obj" "test_compiler.cpp"
>
> I get:
>
> In file included from test_compiler.cpp:1:
> In file included from C:\Utilities\mingw-w64\i686-7.
> 1.0-posix-dwarf-rt_v5-rev0\mingw32\i686-w64-mingw32\include\windows.h:69:
> In file included from C:\Utilities\mingw-w64\i686-7.
> 1.0-posix-dwarf-rt_v5-rev0\mingw32\i686-w64-mingw32\include\windef.h:8:
> In file included from C:\Utilities\mingw-w64\i686-7.
> 1.0-posix-dwarf-rt_v5-rev0\mingw32\i686-w64-mingw32\include\
> minwindef.h:163:
> C:\Utilities\mingw-w64\i686-7.1.0-posix-dwarf-rt_v5-rev0\min
> gw32\i686-w64-mingw32\include\winnt.h:4720:18: error: default
> initialization of an object of const type 'const GUID' (aka 'const _GUID')
> without a user-provided default constructor
>     DEFINE_GUID (GUID_MAX_POWER_SAVINGS, 0xa1841308, 0x3541, 0x4fab, 0xbc,
> 0x81, 0xf7, 0x15, 0x56, 0xf2, 0x0b, 0x4a);
>                  ^
> C:\Utilities\mingw-w64\i686-7.1.0-posix-dwarf-rt_v5-rev0\min
> gw32\i686-w64-mingw32\include\winnt.h:4721:18: error: default
> initialization of an object of const type 'const GUID' (aka 'const _GUID')
> without a user-provided default constructor
>     DEFINE_GUID (GUID_MIN_POWER_SAVINGS, 0x8c5e7fda, 0xe8bf, 0x4a96, 0x9a,
> 0x85, 0xa6, 0xe2, 0x3a, 0x8c, 0x63, 0x5c);
>                  ^
>
> etc. with the same errors for all the other DEFINE_GUID s in winnt.h
>
> Needless to say compiling the same source with mingw-64/gcc-7.1 directly
> succeeds.
>
> Is this a bug in the latest clang source ?
>
>
>
>
>
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20170507/403a7a02/attachment.html>