Merge pull request #100 from mailgun/sergey/restrict

do not parse html quotations if html is longer then certain threshold

setup.py

@@ -29,7 +29,7 @@ class InstallCommand(install):
 
 
 setup(name='talon',
-      version='1.2.12',
+      version='1.2.14',
       description=("Mailgun library "
                    "to extract message quotations and signatures."),
       long_description=open("README.rst").read(),

talon/quotations.py

@@ -164,6 +164,9 @@ RE_PARENTHESIS_LINK = re.compile("\(https?://")
 
 SPLITTER_MAX_LINES = 4
 MAX_LINES_COUNT = 1000
+# an extensive research shows that exceeding this limit
+# leads to excessive processing time
+MAX_HTML_LEN = 2794202
 
 QUOT_PATTERN = re.compile('^>+ ?')
 NO_QUOT_LINE = re.compile('^[^>].*[\S].*')
@@ -382,6 +385,9 @@ def _extract_from_html(msg_body):
     then checking deleted checkpoints,
     then deleting necessary tags.
     """
+    if len(msg_body) > MAX_HTML_LEN:
+        return msg_body
+
     if msg_body.strip() == b'':
         return msg_body
 

tests/html_quotations_test.py

@@ -380,3 +380,15 @@ def test_gmail_forwarded_msg():
 </div><br></div>"""
     extracted = quotations.extract_from_html(msg_body)
     eq_(RE_WHITESPACE.sub('', msg_body), RE_WHITESPACE.sub('', extracted))
+
+
+@patch.object(quotations, 'MAX_HTML_LEN', 1)
+def test_too_large_html():
+    msg_body = 'Reply' \
+               '<div class="gmail_quote">' \
+               '<div class="gmail_quote">On 11-Apr-2011, at 6:54 PM, Bob &lt;bob@example.com&gt; wrote:' \
+               '<div>Test</div>' \
+               '</div>' \
+               '</div>'
+    eq_(RE_WHITESPACE.sub('', msg_body),
+        RE_WHITESPACE.sub('', quotations.extract_from_html(msg_body)))