def test_remove_next_part_from_content(self):
with open(get_test_file("pipermail_nextpart.txt")) as email_file:
msg = message_from_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
contents = scrubber.scrub()[0]
self.assertFalse("-------------- next part --------------" in contents)
def test_remove_next_part_from_content(self):
with open(get_test_file("pipermail_nextpart.txt")) as email_file:
msg = message_from_file(email_file)
scrubber = Scrubber(msg)
contents = scrubber.scrub()[0]
self.failIf("-------------- next part --------------" in contents)
def test_html_only_email(self):
# This email only has an HTML part, thus the scrubbed content will be
# empty. It should be an unicode empty string, not str.
with open(get_test_file("html-email-2.txt")) as email_file:
msg = message_from_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
contents = scrubber.scrub()[0]
self.assertTrue(isinstance(contents, str),
"Scrubbed content should always be unicode")
def test_name_unicode(self):
for num in range(1, 6):
with open(get_test_file("attachment-%d.txt" % num)) as email_file:
msg = message_from_file(email_file)
scrubber = Scrubber(msg)
attachments = scrubber.scrub()[1]
for attachment in attachments:
name = attachment[1]
self.assertTrue(isinstance(name, unicode),
"attachment %r must be unicode" % name)
def _test_non_ascii_payload(self, enc):
with open(get_test_file("payload-%s.txt" % enc), 'rb') as email_file:
msg = message_from_binary_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
contents = scrubber.scrub()[0]
self.assertTrue(isinstance(contents, str))
self.assertEqual(
contents,
'This message contains non-ascii characters:\n\xe9 \xe8 \xe7 \xe0 \xee \xef \xeb \u20ac\n' # noqa
)
def test_bad_content_type(self):
"""Scrubber must handle unknown content-types"""
with open(get_test_file("payload-unknown.txt")) as email_file:
msg = message_from_file(email_file)
scrubber = Scrubber(msg)
try:
contents = scrubber.scrub()[0]
except LookupError, e:
import traceback
print(traceback.format_exc())
self.fail(e) # codec not found
def test_non_ascii_payload(self):
"""Scrubber must handle non-ascii messages"""
for enc in ["utf8", "iso8859"]:
with open(get_test_file("payload-%s.txt" % enc)) as email_file:
msg = message_from_file(email_file)
scrubber = Scrubber(msg)
contents = scrubber.scrub()[0]
self.assertTrue(isinstance(contents, unicode))
self.assertEqual(
contents,
u'This message contains non-ascii '
u'characters:\n\xe9 \xe8 \xe7 \xe0 \xee \xef \xeb \u20ac\n')
def test_attachment_1(self):
with open(get_test_file("attachment-1.txt")) as email_file:
msg = message_from_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
contents, attachments = scrubber.scrub()
self.assertEqual(len(attachments), 1)
self.assertEqual(attachments[0],
(2, 'puntogil.vcf', 'text/x-vcard', "utf-8",
'begin:vcard\nfn:gil\nn:;gil\nversion:2.1\n'
'end:vcard\n\n'))
self.assertEqual(
contents, "This is a test message.\n\n"
"\n-- \ndevel mailing list\[email protected]\n"
"https://admin.fedoraproject.org/mailman/listinfo/devel\n")
def test_html_email_1(self):
with open(get_test_file("html-email-1.txt")) as email_file:
msg = message_from_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
contents, attachments = scrubber.scrub()
self.assertEqual(len(attachments), 1)
# HTML part
self._check_html_attachment(
attachments[0], (2, "attachment.html", "text/html", "iso-8859-1"))
self.assertEqual(len(attachments[0][4]), 2688)
# Scrubbed content
self.assertEqual(
contents, "This is a test message\n"
"Non-ASCII chars: r\xe9ponse fran\xe7ais \n\n\n")
def test_attachment_5(self):
with open(get_test_file("attachment-5.txt")) as email_file:
msg = message_from_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
contents, attachments = scrubber.scrub()
self.assertEqual(len(attachments), 1)
# text attachment
self.assertEqual(attachments[0][0:4],
(2, "todo-déjeuner.txt", "text/plain", "utf-8"))
self.assertEqual(len(attachments[0][4]), 110)
# Scrubbed content
self.assertEqual(
contents, 'This is a test, HTML message with '
'accented letters : \xe9 \xe8 \xe7 \xe0.\nAnd an '
'attachment with an accented filename\n\n\n\n\n\n')
def test_attachment_name_badly_encoded(self):
msg = Message()
msg["From"] = "*****@*****.**"
msg["Message-ID"] = "<dummy>"
msg.set_payload(b"Dummy content")
msg.add_header(b'Content-Disposition', b'attachment',
filename=b'non-ascii-\xb8\xb1\xb1\xbe.jpg')
scrubber = Scrubber(msg)
try:
attachments = scrubber.scrub()[1]
except UnicodeDecodeError:
print(format_exc())
self.fail("Could not decode the filename")
self.assertEqual(
attachments,
[(0, u'attachment.bin', 'text/plain', None, b'Dummy content')])
def test_attachment_3(self):
with open(get_test_file("attachment-3.txt")) as email_file:
msg = message_from_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
contents, attachments = scrubber.scrub()
self.assertEqual(len(attachments), 2)
# HTML part
self._check_html_attachment(
attachments[0], (3, "attachment.html", "text/html", "iso-8859-1"))
self.assertEqual(len(attachments[0][4]), 3114)
# Image attachment
self.assertEqual(attachments[1][0:4],
(4, "GeoffreyRoucourt.jpg", "image/jpeg", None))
self.assertEqual(len(attachments[1][4]), 282180)
# Scrubbed content
self.assertEqual(contents, "This is a test message\n\n\n")
def test_attachment_name_badly_encoded(self):
with open(get_test_file("email-bad-filename.txt"), 'rb') as email_file:
msg = message_from_binary_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
try:
attachments = scrubber.scrub()[1]
except UnicodeDecodeError:
print(format_exc())
self.fail("Could not decode the filename")
# The filename has non-ascii characters without the encoding specified,
# Python will try to decode their name with best guess (ascii) and then
# replace the characters that don't correspond to an ascii code
# point. Then, we scrub the filename to allow only alpahun with dash,
# underscore and dot.
self.assertEqual(
attachments,
[(0, 'non-ascii-u3b5.jpg', 'text/plain', None, 'Dummy content\n')])
def test_attachment_2(self):
with open(get_test_file("attachment-2.txt")) as email_file:
msg = message_from_file(email_file)
scrubber = Scrubber(msg)
contents, attachments = scrubber.scrub()
self.assertEqual(len(attachments), 1)
self.assertEqual(attachments[0], (
3, u'signature.asc', u'application/pgp-signature', None,
'-----BEGIN PGP SIGNATURE-----\r\nVersion: GnuPG v1.4.12 '
'(GNU/Linux)\r\nComment: Using GnuPG with Mozilla - '
'http://www.enigmail.net/\r\n\r\niEYEARECAAYFAlBhm3oACgkQhmBj'
'z394AnmMnQCcC+6tWcqE1dPQmIdRbLXgKGVp\r\nEeUAn2OqtaXaXaQV7rx+'
'SmOldmSzcFw4\r\n=OEJv\r\n-----END PGP SIGNATURE-----\r\n'))
self.assertEqual(
contents,
u"This is a test message\r\nNon-ascii chars: Hofm\xfchlgasse\r\n"
u"\n-- \ndevel mailing list\[email protected]\n"
u"https://admin.fedoraproject.org/mailman/listinfo/devel\n"
)
def test_attachment_4(self):
with open(get_test_file("attachment-4.txt")) as email_file:
msg = message_from_file(email_file, policy=policy.SMTP)
scrubber = Scrubber(msg)
contents, attachments = scrubber.scrub()
self.assertEqual(len(attachments), 2)
# HTML part
self._check_html_attachment(
attachments[0], (3, "attachment.html", "text/html", "iso-8859-1"))
self.assertEqual(len(attachments[0][4]), 113)
# text attachment
self.assertEqual(attachments[1][0:4],
(4, "todo-déjeuner.txt", "text/plain", "utf-8"))
self.assertEqual(len(attachments[1][4]), 110)
# Scrubbed content
self.assertEqual(
contents, 'This is a test, HTML message with '
'accented letters : \xe9 \xe8 \xe7 \xe0.\nAnd an '
'attachment with an accented filename\n\n\n\n\n')
def parse(message):
msgobj = EmailParser().parsestr(message)
header_parser = HeaderParser()
headers = []
headers_dict = header_parser.parsestr(message)
for key in headers_dict.keys():
headers += ['{}: {}'.format(key, headers_dict[key])]
content = Scrubber(msgobj).scrub()[0]
return {
'body': content,
'headers': '\n'.join(headers),
}
def parse(message):
msgobj = message_from_string(message, policy=policy.SMTP)
header_parser = HeaderParser()
headers = []
headers_dict = header_parser.parsestr(message)
for key in headers_dict.keys():
headers.append('{}: {}'.format(key, headers_dict[key]))
content = Scrubber(msgobj).scrub()[0]
return {
'body': content,
'headers': '\n'.join(headers),
}
def parse(message):
msgobj = EmailParser().parsestr(message)
header_parser = HeaderParser()
if msgobj['Subject'] is not None:
decodefrag = decode_header(msgobj['Subject'])
subj_fragments = []
for s, enc in decodefrag:
if enc:
s = unicode(s, enc).encode('utf8', 'replace')
subj_fragments.append(s)
subject = ''.join(subj_fragments)
else:
subject = None
headers = []
headers_dict = header_parser.parsestr(message)
for key in headers_dict.keys():
headers += ['{}: {}'.format(key, headers_dict[key])]
content = Scrubber(msgobj).scrub()[0]
return {
'subject': subject,
'body': content,
'headers': '\n'.join(headers),
}
def add_to_list(list_name, message):
# timeit("1 start")
mlist = MailingList.objects.get_or_create(name=list_name)[0]
if not getattr(settings, "HYPERKITTY_BATCH_MODE", False):
mlist.update_from_mailman()
mlist.save()
if mlist.archive_policy == ArchivePolicy.never.value:
logger.info("Archiving disabled by list policy for %s", list_name)
return
if "Message-Id" not in message:
raise ValueError("No 'Message-Id' header in email", message)
# timeit("2 after ml, before checking email & sender")
msg_id = get_message_id(message)
if Email.objects.filter(mailinglist=mlist, message_id=msg_id).exists():
raise DuplicateMessage(msg_id)
email = Email(mailinglist=mlist, message_id=msg_id)
email.in_reply_to = get_ref(message) # Find thread id
if message.get_unixfrom() is not None:
mo = UNIXFROM_DATE_RE.match(message.get_unixfrom())
if mo:
archived_date = parsedate(mo.group(1))
if archived_date is not None:
email.archived_date = archived_date
# Sender
try:
from_name, from_email = parseaddr(message['From'])
from_name = header_to_unicode(from_name).strip()
sender_address = from_email.decode("ascii").strip()
except (UnicodeDecodeError, UnicodeEncodeError):
raise ValueError("Non-ascii sender address", message)
if not sender_address:
if from_name:
sender_address = re.sub("[^a-z0-9]", "", from_name.lower())
if not sender_address:
sender_address = "unknown"
sender_address = "{}@example.com".format(sender_address)
else:
sender_address = "*****@*****.**"
email.sender_name = from_name
sender = Sender.objects.get_or_create(address=sender_address)[0]
email.sender = sender
if not getattr(settings, "HYPERKITTY_BATCH_MODE", False):
try:
sender.set_mailman_id()
except MailmanConnectionError:
pass
# timeit("3 after sender, before email content")
# Headers
email.subject = header_to_unicode(message.get('Subject'))
if email.subject is not None:
# limit subject size to 512, it's a varchar field
email.subject = email.subject[:512]
msg_date = parsedate(message.get("Date"))
if msg_date is None:
# Absent or unparseable date
msg_date = timezone.now()
utcoffset = msg_date.utcoffset()
if msg_date.tzinfo is not None:
msg_date = msg_date.astimezone(timezone.utc) # store in UTC
email.date = msg_date
if utcoffset is None:
email.timezone = 0
else:
# in minutes
email.timezone = int(
((utcoffset.days * 24 * 60 * 60) + utcoffset.seconds) / 60)
# Content
scrubber = Scrubber(message)
# warning: scrubbing modifies the msg in-place
email.content, attachments = scrubber.scrub()
# timeit("4 after email content, before signals")
# TODO: detect category?
# Set or create the Thread
if email.in_reply_to is not None:
try:
ref_msg = Email.objects.get(mailinglist=email.mailinglist,
message_id=email.in_reply_to)
except Email.DoesNotExist:
# the parent may not be archived (on partial imports), create a new
# thread for now.
pass
else:
# re-use parent's thread-id
email.parent = ref_msg
email.thread_id = ref_msg.thread_id
thread = ref_msg.thread
thread_created = False
if email.thread_id is None:
# Create the thread if not found
thread, thread_created = Thread.objects.get_or_create(
mailinglist=email.mailinglist, thread_id=email.message_id_hash)
email.thread = thread
email.save() # must save before setting the thread.starting_email
thread.date_active = email.date
if thread_created:
thread.starting_email = email
thread.save()
if thread_created:
new_thread.send("Mailman", thread=thread)
# signal_results = new_thread.send_robust("Mailman", thread=thread)
# for receiver, result in signal_results:
# if isinstance(result, Exception):
# logger.warning(
# "Signal 'new_thread' to {} raised an "
# "exception: {}".format(
# receiver.func_name, result))
# Signals
new_email.send("Mailman", email=email)
# signal_results = new_email.send_robust("Mailman", email=email)
# for receiver, result in signal_results:
# if isinstance(result, Exception):
# logger.warning(
# "Signal 'new_email' to {} raised an exception: {}".format(
# receiver.func_name, result))
# #logger.exception(result)
# #from traceback import print_exc; print_exc(result)
# timeit("5 after signals, before save")
# timeit("6 after save")
# compute thread props here because email must have been saved before
# (there will be DB queries in this function)
if not getattr(settings, "HYPERKITTY_BATCH_MODE", False):
compute_thread_order_and_depth(email.thread)
# Attachments (email must have been saved before)
for attachment in attachments:
counter, name, content_type, encoding, content = attachment
if Attachment.objects.filter(email=email, counter=counter).exists():
continue
Attachment.objects.create(email=email,
counter=counter,
name=name,
content_type=content_type,
encoding=encoding,
content=content)
return email.message_id_hash
def get_attachments(message):
message = EmailParser().parsestr(message)
return Scrubber(message).scrub()[1]
def add_to_list(list_name, message):
assert isinstance(message, EmailMessage)
# timeit("1 start")
mlist = MailingList.objects.get_or_create(name=list_name)[0]
if not getattr(settings, "HYPERKITTY_BATCH_MODE", False):
update_from_mailman(mlist.name)
mlist.save()
if mlist.archive_policy == ArchivePolicy.never.value:
logger.info("Archiving disabled by list policy for %s", list_name)
return
if "Message-Id" not in message:
raise ValueError("No 'Message-Id' header in email", message)
# timeit("2 after ml, before checking email & sender")
msg_id = get_message_id(message)
if Email.objects.filter(mailinglist=mlist, message_id=msg_id).exists():
raise DuplicateMessage(msg_id)
email = Email(mailinglist=mlist, message_id=msg_id)
email.in_reply_to = get_ref(message) # Find thread id
if message.get_unixfrom() is not None:
mo = UNIXFROM_DATE_RE.match(message.get_unixfrom())
if mo:
archived_date = parsedate(mo.group(1))
if archived_date is not None:
email.archived_date = archived_date
# Sender
try:
from_str = header_to_unicode(message['From'])
from_name, from_email = parseaddr(from_str)
from_name = from_name.strip()
sender_address = from_email.encode('ascii').decode("ascii").strip()
except (UnicodeDecodeError, UnicodeEncodeError):
raise ValueError("Non-ascii sender address", message)
if not sender_address:
if from_name:
sender_address = re.sub("[^a-z0-9]", "", from_name.lower())
if not sender_address:
sender_address = "unknown"
sender_address = "{}@example.com".format(sender_address)
else:
sender_address = "*****@*****.**"
email.sender_name = from_name
sender = Sender.objects.get_or_create(address=sender_address)[0]
email.sender = sender
if not getattr(settings, "HYPERKITTY_BATCH_MODE", False):
sender_mailman_id(sender.pk)
# timeit("3 after sender, before email content")
# Headers
email.subject = header_to_unicode(message.get('Subject'))
if email.subject is not None:
# limit subject size to 512, it's a varchar field
email.subject = email.subject[:512]
msg_date = parsedate(message.get("Date"))
if msg_date is None:
# Absent or unparseable date
msg_date = timezone.now()
utcoffset = msg_date.utcoffset()
if msg_date.tzinfo is not None:
msg_date = msg_date.astimezone(timezone.utc) # store in UTC
email.date = msg_date
if utcoffset is None:
email.timezone = 0
else:
# in minutes
email.timezone = int(
((utcoffset.days * 24 * 60 * 60) + utcoffset.seconds) / 60)
# Content
scrubber = Scrubber(message)
# warning: scrubbing modifies the msg in-place
email.content, attachments = scrubber.scrub()
# timeit("4 after email content, before signals")
# TODO: detect category?
# Find the parent email.
# This can't be moved to Email.on_pre_save() because Email.set_parent()
# needs to be free to change the parent independently from the in_reply_to
# property, and will save() the instance.
# This, along with some of the work done in Email.on_pre_save(), could be
# moved to an async task, but the rest of the app must be able to cope with
# emails lacking this data, and email being process randomly (child before
# parent). The work in Email.on_post_created() also depends on it, so be
# careful with task dependencies if you ever do this.
# Plus, it has "premature optimization" written all over it.
if email.in_reply_to is not None:
try:
ref_msg = Email.objects.get(mailinglist=email.mailinglist,
message_id=email.in_reply_to)
except Email.DoesNotExist:
# the parent may not be archived (on partial imports), create a new
# thread for now.
pass
else:
# re-use parent's thread-id
email.parent = ref_msg
email.thread_id = ref_msg.thread_id
try:
email.save()
except DataError as e:
raise ValueError(str(e))
# Attachments (email must have been saved before)
for attachment in attachments:
counter, name, content_type, encoding, content = attachment
if Attachment.objects.filter(email=email, counter=counter).exists():
continue
att = Attachment.objects.create(email=email,
counter=counter,
name=name,
content_type=content_type,
encoding=encoding)
att.set_content(content)
att.save()
return email.message_id_hash
def get_attachments(message):
message = message_from_string(message, policy=policy.SMTP)
return Scrubber(message).scrub()[1]