PATCH:kernel32/locale:Fix some bugs about special Chinese test in parse_locale_name()

[Nikolay Sivov]

On 14.04.2017 13:21, THuang wrote:
> 
> 


> @@ -457,6 +457,8 @@ static void parse_locale_name( const WCHAR *str, struct locale_name *name )
>      static const WCHAR cW[] = {'C',0};
>      static const WCHAR latinW[] = {'l','a','t','i','n',0};
>      static const WCHAR latnW[] = {'-','L','a','t','n',0};
> +    static const WCHAR HantW[] = {'-','H','a','n','t',0};
> +    static const WCHAR HansW[] = {'-','H','a','n','s',0};
>      WCHAR *p;
>  
>      TRACE("%s\n", debugstr_w(str));
> @@ -487,6 +489,16 @@ static void parse_locale_name( const WCHAR *str, struct locale_name *name )
>      else if (*p == '-')  /* Windows format */
>      {
>          strcpyW( name->win_name, name->lang );
> +        if(!strcmpiW(p,HantW)){
> +            name->lcid = 0x0c04;
> +            name->matches = 1;
> +            return ;
> +        }
> +        if(!strcmpiW(p,HansW)){
> +            name->lcid = 0x0804;
> +            name->matches = 1;
> +            return ;
> +        }
>          *p++ = 0;
>          name->country = p;
>          if (!(p = strpbrkW( p, winsepW ))) goto done;

Well, no, that's not how it works. Did you explore a possibility if
adding hant/hans specific locale data to kernel32/nls will help?