From 3172795aa20cadbcc0507f0f40019c779173580b Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Mon, 24 Mar 2008 07:29:07 +0000
Subject: [PATCH] Fix index parsing of non-ascii web sites

---
 src/libprs500/web/feeds/news.py | 7 +++----
 1 file changed, 3 insertions(+), 4 deletions(-)

diff --git a/src/libprs500/web/feeds/news.py b/src/libprs500/web/feeds/news.py
index f1e8f3b9f0..68ca49d5ac 100644
--- a/src/libprs500/web/feeds/news.py
+++ b/src/libprs500/web/feeds/news.py
@@ -265,10 +265,9 @@ class BasicNewsRecipe(object):
             raw = url_or_raw
         if not isinstance(raw, unicode) and self.encoding:
             raw = raw.decode(self.encoding)
-        raw = re.sub(r'&(\S+?);', 
-                     lambda match: entity_to_unicode(match, encoding=self.encoding), 
-                     raw)
-        return BeautifulSoup(raw)
+        massage = list(BeautifulSoup.MARKUP_MASSAGE)
+        massage.append((re.compile(r'&(\S+?);'), lambda match: entity_to_unicode(match, encoding=self.encoding))) 
+        return BeautifulSoup(raw, markupMassage=massage)
         
     
     def sort_index_by(self, index, weights):