Removing excess carriage returns

muggot · Feb 17, 2013 · 430116f · 430116f
1 parent 7420632
commit 430116f
Showing 1 changed file with 6 additions and 2 deletions.
diff --git a/goose/outputformatters.py b/goose/outputformatters.py
@@ -68,8 +68,12 @@ def convertToText(self):
             if txt:
                 txt = HTMLParser().unescape(txt)
                 txts.append(innerTrim(txt))
-        text = '\n\n'.join(txts)
-	text = re.sub(u'[\ufffc]','\n\n',text)
+        text = '\n'.join(txts)
+	text = re.sub(u'[\ufffc]','\n',text)
+        lines = text.split('\n')
+        text = ''
+        for line in lines:
+            if re.search('[^ \t\r]',line): text += line + '\n'
         return text
 
     def convertLinksToText(self):