Re: msvcrt: Partially implement type_info::name_internal_method()

[Alexandre Julliard]

Nikolay Sivov  writes:
> @@ -260,3 +307,24 @@ int CDECL _atoflt( _CRT_FLOAT *value, char *str )
>  {
>      return _atoflt_l( value, str, NULL );
>  }
> -
>
> +/*********************************************************************
> -
> - ?_name_internal_method@type_info@@QBEPBDPAU__type_info_node@@@Z (MSVCR90.@)
> - */
>
> +DEFINE_THISCALL_WRAPPER(MSVCRT_type_info_name_internal_method,8)
> +const char * __thiscall MSVCRT_type_info_name_internal_method(type_info * _this, struct __type_info_node *node)
> +{
> - /* this is a forward to type_info::name() */
>
> +#ifdef __i386__
> - static const char *typeinfo_name = "?name@type_info@@QBEPBDXZ";
> - void* (WINAPI *p)(void*);
>
> +#else
> - static const char typeinfo_name[] = "?name@type_info@@QEBAPEBDXZ";
> - void* (__cdecl *p)(void*);
>
> +#endif
> - static int once;
> -
> - if (node && !once++) FIXME("type_info_node parameter ignored\n");
> - p = (void*)GetProcAddress(GetModuleHandleA("msvcrt.dll"), typeinfo_name);
> - return call_func1(p, _this);
>
> +}

I think it would be better to duplicate the implementation after
all. Sorry for misleading you with the GetProcAddress suggestion,
I hadn't realized that it would require the thiscall mess.