From 4ee46c0a977477e889209f3f5940f1f759f5bd5d Mon Sep 17 00:00:00 2001
From: Sergey Obukhov <sergey.obykhov@mailgunhq.com>
Date: Tue, 9 Aug 2016 17:08:58 -0700
Subject: [PATCH 1/3] do not parse html quotations if html is longer then
 certain threshold

---
 talon/quotations.py | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/talon/quotations.py b/talon/quotations.py
index d9dba21..f6122ff 100644
--- a/talon/quotations.py
+++ b/talon/quotations.py
@@ -164,6 +164,7 @@ RE_PARENTHESIS_LINK = re.compile("\(https?://")
 
 SPLITTER_MAX_LINES = 4
 MAX_LINES_COUNT = 1000
+MAX_HTML_LEN = 2794202
 
 QUOT_PATTERN = re.compile('^>+ ?')
 NO_QUOT_LINE = re.compile('^[^>].*[\S].*')
@@ -382,6 +383,9 @@ 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
 

From 21e9a31ffe981ae9d58b4143b478b89232875d8d Mon Sep 17 00:00:00 2001
From: Sergey Obukhov <sergey.obykhov@mailgunhq.com>
Date: Tue, 9 Aug 2016 17:15:49 -0700
Subject: [PATCH 2/3] add test

---
 tests/html_quotations_test.py | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/tests/html_quotations_test.py b/tests/html_quotations_test.py
index c155cbb..03c66a8 100644
--- a/tests/html_quotations_test.py
+++ b/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)))

From a0d7236d0b8d7c2635d0aff040a485b8b0365d36 Mon Sep 17 00:00:00 2001
From: Sergey Obukhov <sergey.obykhov@mailgunhq.com>
Date: Thu, 11 Aug 2016 15:49:09 -0700
Subject: [PATCH 3/3] bump version and add a comment

---
 setup.py            | 2 +-
 talon/quotations.py | 2 ++
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/setup.py b/setup.py
index 83f0714..8f054df 100755
--- a/setup.py
+++ b/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(),
diff --git a/talon/quotations.py b/talon/quotations.py
index f6122ff..9999e6a 100644
--- a/talon/quotations.py
+++ b/talon/quotations.py
@@ -164,6 +164,8 @@ 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('^>+ ?')