Flanker barfs on malformed MIME parts

[khamidou]

Hi,

We use flanker to parse MIME parts and I think I found a special case where the parser crashes on slightly-malformed content. I've narrowed it to the following test case:

Delivered-To: [email protected]
Date: 11 Jan 2013 18:54:26 -0000
MIME-Version: 1.0
To: <[email protected]>
Message-ID: <1357884894.S.69618.18751.f5mail-224-118.example.com>
Sender: [email protected]
Subject: Dear sir
From: "John Doe " <[email protected]>
Content-Type: multipart/mixed;
    boundary="=_e6ddd3579a993208589b263b76d66bec"

--=_e6ddd3579a993208589b263b76d66bec
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset="UTF-8"

URGENT - HELP ME DISTRIBUTE MY $15 MILLION TO CHARITY

IN SUMMARY:- I have 15,000,000.00 (fifteen million) U.S. Dollars and I want you to assist me in distributing the money to charity organizations.

--=_e6ddd3579a993208589b263b76d66bec
Content-Transfer-Encoding: 
Content-Type: message/rfc822;
 name="ForwardedMessage"; 
Content-Disposition: inline;
 filename="ForwardedMessage"; 
--=_e6ddd3579a993208589b263b76d66bec--

I think the parser expects the Message part to be followed by \r\n which makes it crash.

I use the following program to trigger the bug:

import sys
from flanker import mime

fd = open(sys.argv[1], "r")
contents = fd.read()

parsed = mime.from_string(contents)

for mimepart in parsed.walk(with_self=parsed.content_type.is_singlepart()):
    print mimepart.headers

The traceback is:

Traceback (most recent call last):
  File "/contrib/flanker/testcase.py", line 11, in <module>
    print mimepart.headers
  File "/contrib/flanker/flanker/mime/message/part.py", line 389, in headers
    return self._container.headers
  File "/contrib/flanker/flanker/mime/message/part.py", line 42, in headers
    self._load_headers()
  File "/contrib/flanker/flanker/mime/message/part.py", line 65, in _load_headers
    self.stream.seek(self.start)
TypeError: an integer is required

I'd be happy to contribute a patch if you could point me in the right direction.

[grinich]

@khamidou Can you verify the above PR fixes the issue for you as well?