From 9a08f7feb8bb797932fddec51e096ee3600cafd4 Mon Sep 17 00:00:00 2001
From: Miroslav Stampar <miroslav.stampar@gmail.com>
Date: Tue, 12 Oct 2010 20:01:59 +0000
Subject: [PATCH] minor update

---
 lib/core/common.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lib/core/common.py b/lib/core/common.py
index b9564207f..0c3e05717 100644
--- a/lib/core/common.py
+++ b/lib/core/common.py
@@ -1128,7 +1128,9 @@ def preparePageForLineComparison(page):
 def getFilteredPageContent(page):
     retVal = page
     if isinstance(page, basestring):
-        retVal = re.sub(r"(?s)<script.+?</script>|<style.+?</style>|<[^>]+>|\t|\n|\r", "", page)
+        retVal = re.sub(r"(?s)<script.+?</script>|<style.+?</style>|<[^>]+>|\t|\n|\r", " ", page)
+        while retVal.find("  ") != -1:
+            retVal = retVal.replace("  ", " ")
     return retVal
 
 def getPageTextWordsSet(page):