Strings were stored in the Lexicon using the intern() function

to provide some minor speedup. But this made it impossible to
store unicode strings. So we can now store unicode strings
in a Lexicon *yippi*

lib/python/Products/PluginIndexes/TextIndex/Lexicon.py

@@ -98,6 +98,7 @@ from BTrees.IOBTree import IOBTree
 from BTrees.IIBTree import IISet, IITreeSet
 
 from randid import randid
+from types import StringType
 
 class Lexicon(Persistent, Implicit):
     """Maps words to word ids and then some
@@ -198,7 +199,11 @@ class Lexicon(Persistent, Implicit):
         while not inverse.insert(wid, word):
             wid=randid()
 
+        if isinstance(word,StringType):        
             self._lexicon[intern(word)] = wid
+        else:
+            self._lexicon[word] = wid
+            
 
         return wid