BUG# 10687: Merge engine fails under Windows

This cset is to roll out the cset applied earlier from Ingo.
This bug has been fixed with a different cset.



sql/ha_myisammrg.cc:
  removed the patch applied earlier
strings/my_vsnprintf.c:
  removed the patch applied earlier.  The patch was to add %c ability.  
  Monty has already done this.

sql/ha_myisammrg.cc

@@ -406,8 +406,8 @@ int ha_myisammrg::create(const char *name, register TABLE *form,
         This means that it might not be possible to move the DATADIR of
         an embedded server without changing the paths in the .MRG file.
       */
-      uint length= my_snprintf(buff, FN_REFLEN, "%s%c%s/%s", mysql_data_home,
-			       FN_LIBCHAR, tables->db, tables->real_name);
+      uint length= my_snprintf(buff, FN_REFLEN, "%s%s/%s", mysql_data_home,
+			       tables->db, tables->real_name);
       /*
         If a MyISAM table is in the same directory as the MERGE table,
         we use the table name without a path. This means that the

strings/my_vsnprintf.c

@@ -28,7 +28,6 @@
     %#[l]u
     %#[l]x
     %#.#s	Note first # is ignored
-    %c
     
   RETURN
     length of result string
@@ -121,11 +120,6 @@ int my_vsnprintf(char *to, size_t n, const char* fmt, va_list ap)
       to+= res_length;
       continue;
     }
-    else if (*fmt == 'c')
-    {
-      *(to++)= (char) va_arg(ap, int);
-      continue;
-    }
     /* We come here on '%%', unknown code or too long parameter */
     if (to == end)
       break;