Improved table-by-name recognition for non-latin window titles

What i mean:
  u'1464739' in 'Speed #1464739 -  \xc1\xcb \xd5\xee\xeb\xe4\xe5\xec' -> Exception
  u'1464739' in 'Speed #1464739 -  \xc1\xcb \xd5\xee\xeb\xe4\xe5\xec'.decode('cp1251') -> True

pyfpdb/Tables.py

@@ -230,11 +230,20 @@ def discover_nt_by_name(c, tablename):
     """Finds poker client window with the given table name."""
     titles = {}
     win32gui.EnumWindows(win_enum_handler, titles)
+    
+    def getDefaultEncoding():
+        # FIXME: if somebody know better place fot this function - move it
+        # FIXME: it's better to use GetCPInfo for windows http://msdn.microsoft.com/en-us/library/dd318078(VS.85).aspx
+        # but i have no idea, how to call it
+        import locale
+        return locale.getpreferredencoding()
+        
     for hwnd in titles:
         #print "Tables.py: tablename =", tablename, "title =", titles[hwnd]
         try:
+            # maybe it's better to make global titles[hwnd] decoding?
             # this can blow up in XP on some windows, eg firefox displaying http://docs.python.org/tutorial/classes.html
-            if not tablename.lower() in titles[hwnd].lower(): continue
+            if not tablename.lower() in titles[hwnd].decode(getDefaultEncoding()).lower(): continue
         except:
             continue
         if 'History for table:' in titles[hwnd]: continue # Everleaf Network HH viewer window