user32: Support utf-8 in WM_CHAR W->A conversion.

dlls/user32/message.c

@@ -183,9 +183,9 @@ BOOL map_wparam_AtoW( UINT message, WPARAM *wparam, enum wm_char_mapping mapping
  */
 static void map_wparam_WtoA( MSG *msg, BOOL remove )
 {
-    BYTE ch[4];
+    BYTE ch[4] = { 0 };
     WCHAR wch[2];
-    DWORD len;
+    DWORD i, len;
     DWORD cp;
 
     switch(msg->message)
@@ -195,7 +195,6 @@ static void map_wparam_WtoA( MSG *msg, BOOL remove )
         {
             cp = get_input_codepage();
             wch[0] = LOWORD(msg->wParam);
-            ch[0] = ch[1] = 0;
             len = WideCharToMultiByte( cp, 0, wch, 1, (LPSTR)ch, 2, NULL, NULL );
             if (len == 2)  /* DBCS char */
             {
@@ -224,14 +223,12 @@ static void map_wparam_WtoA( MSG *msg, BOOL remove )
         cp = get_input_codepage();
         wch[0] = LOWORD(msg->wParam);
         wch[1] = HIWORD(msg->wParam);
-        ch[0] = ch[1] = 0;
-        WideCharToMultiByte( cp, 0, wch, 2, (LPSTR)ch, 4, NULL, NULL );
-        msg->wParam = MAKEWPARAM( ch[0] | (ch[1] << 8), 0 );
+        len = WideCharToMultiByte( cp, 0, wch, 2, (LPSTR)ch, 4, NULL, NULL );
+        for (msg->wParam = i = 0; i < len; i++) msg->wParam |= ch[i] << (8 * i);
         break;
     case WM_IME_CHAR:
         cp = get_input_codepage();
         wch[0] = LOWORD(msg->wParam);
-        ch[0] = ch[1] = 0;
         len = WideCharToMultiByte( cp, 0, wch, 1, (LPSTR)ch, 2, NULL, NULL );
         if (len == 2)
             msg->wParam = MAKEWPARAM( (ch[0] << 8) | ch[1], HIWORD(msg->wParam) );