Re: [Gnash-dev] Development for Win32

[Tonko Juricic]

On Sun, Nov 16, 2008 at 8:27 PM, Bastiaan Jacques <address@hidden> wrote:

> Getting it to build in Visual Studio is probably a fairly time consuming
> task, although our code should be pretty portable (and compiles with
> Ming on win32 as-is). Also, you can ask Dossy win32 specific questions
> since he is the current porter for that platform.
>
> Feel free to ask questions (in private or on mailing lists, whichever
> you prefer).

Indeed, the code is very portable and the first issues I am facing
have to do with the fact that Ming on Win32
keeps several UNIX-specific include files.

I think it is best to try distinguish Ming and MSVC Win32 differences
so that I don't break existing port that works.
One way to know that code is compiled by Microsoft compiler is
_MSC_VER macro which additionally can
specify compiler versions, if it ever becomes required.

I am using this macro in libbase/utility.h file like this:

inline bool isFinite(double d)
{
#if defined(WIN32) && defined(_MSC_VER)
        return _finite(d);
#elif defined(HAVE_FINITE) && !defined(HAVE_ISFINITE)
    return (finite(d));
#else
    // Put using namespace std; here if you have to
    // put it anywhere.
    using namespace std;
    return (isfinite(d));
#endif
}


Also I would like to ask Dossy the question about
libbase/extension.cpp, at the very top there is probably his comment:

#if HAVE_DIRENT_H || WIN32==1    // win32 hack                <-- ?????
# include <dirent.h>
# define NAMLEN(dirent) std::strlen((dirent)->d_name)
#else
# define dirent direct
...

<dirent.h> does not exist in MS VC++ and I don't understand the hack.

Thanks,
Tony