Better cleanup of generic articles and lefigaro

try_drop_tree() was only removing the first result.
closes #725

modules/lefigaro/test.py

@@ -19,6 +19,7 @@
 
 
 from weboob.tools.test import BackendTest
+from weboob.tools.misc import html2text
 
 
 __all__ = ['LeFigaroTest']
@@ -30,3 +31,19 @@ class LeFigaroTest(BackendTest):
     def test_new_messages(self):
         for message in self.backend.iter_unread_messages():
             pass
+
+    def test_content(self):
+        urls = ['http://www.lefigaro.fr/international/2011/10/24/01003-20111024ARTFIG00704-les-islamo-conservateurs-maitres-du-jeu-tunisien.php',
+                'http://www.lefigaro.fr/international/2012/01/29/01003-20120129ARTFIG00191-floride-la-primaire-suspendue-a-l-humeur-des-hispaniques.php']
+
+        for url in urls:
+            thread = self.backend.get_thread(url)
+            assert len(thread.root.content)
+            assert '<script' not in thread.root.content
+            assert 'object' not in thread.root.content
+            assert 'BFM' not in thread.root.content
+
+            assert 'AUSSI' not in thread.root.content
+
+            # no funny tags means html2text does not crash
+            assert len(html2text(thread.root.content))