8000 [3.10] gh-121284: Fix email address header folding with parsed encode… · python/cpython@cfaee20 · GitHub
[go: up one dir, main page]
Skip to content

Commit cfaee20

Browse files
encukoumedmundsbitdancer
authored
[3.10] gh-121284: Fix email address header folding with parsed encoded-word (GH-122754) (GH-131411)
Email generators using email.policy.default may convert an RFC 2047 encoded-word to unencoded form during header refolding. In a structured header, this could allow 'specials' chars outside a quoted-string, leading to invalid address headers and enabling spoofing. This change ensures a parsed encoded-word that contains specials is kept as an encoded-word while the header is refolded. [Better fix from @bitdancer.] (cherry picked from commit 295b53d) Co-authored-by: Mike Edmunds <medmunds@gmail.com> Co-authored-by: R David Murray <rdmurray@bitdance.com>
1 parent 48f455e commit cfaee20

File tree

3 files changed

+37
-5
lines changed

3 files changed

+37
-5
lines changed

‎Lib/email/_header_value_parser.py

Lines changed: 5 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -1037,7 +1037,7 @@ def get_fws(value):
10371037
fws = WhiteSpaceTerminal(value[:len(value)-len(newvalue)], 'fws')
10381038
return fws, newvalue
10391039

1040-
def get_encoded_word(value):
1040+
def get_encoded_word(value, terminal_type='vtext'):
10411041
""" encoded-word = "=?" charset "?" encoding "?" encoded-text "?="
10421042
10431043
"""
@@ -1076,7 +1076,7 @@ def get_encoded_word(value):
10761076
ew.append(token)
10771077
continue
10781078
chars, *remainder = _wsp_splitter(text, 1)
1079-
vtext = ValueTerminal(chars, 'vtext')
1079+
vtext = ValueTerminal(chars, terminal_type)
10801080
_validate_xtext(vtext)
10811081
ew.append(vtext)
10821082
text = ''.join(remainder)
@@ -1118,7 +1118,7 @@ def get_unstructured(value):
11181118
valid_ew = True
11191119
if value.startswith('=?'):
11201120
try:
1121-
token, value = get_encoded_word(value)
1121+
token, value = get_encoded_word(value, 'utext')
11221122
except _InvalidEwError:
11231123
valid_ew = False
11241124
except errors.HeaderParseError:
@@ -1147,7 +1147,7 @@ def get_unstructured(value):
11471147
# the parser to go in an infinite loop.
11481148
if valid_ew and rfc2047_matcher.search(tok):
11491149
tok, *remainder = value.partition('=?')
1150-
vtext = ValueTerminal(tok, 'vtext')
1150+
vtext = ValueTerminal(tok, 'utext')
11511151
_validate_xtext(vtext)
11521152
unstructured.append(vtext)
11531153
value = ''.join(remainder)
@@ -2781,7 +2781,7 @@ def _refold_parse_tree(parse_tree, *, policy):
27812781
continue
27822782
tstr = str(part)
27832783
if not want_encoding:
2784-
if part.token_type == 'ptext':
2784+
if part.token_type in ('ptext', 'vtext'):
27852785
# Encode if tstr contains special characters.
27862786
want_encoding = not SPECIALSNL.isdisjoint(tstr)
27872787
else:

‎Lib/test/test_email/test__header_value_parser.py

Lines changed: 25 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -2946,6 +2946,31 @@ def test_address_list_with_unicode_names_in_quotes(self):
29462946
'=?utf-8?q?H=C3=BCbsch?= Kaktus <beautiful@example.com>,\n'
29472947
' =?utf-8?q?bei=C3=9Ft_bei=C3=9Ft?= <biter@example.com>\n')
29482948

2949+
def test_address_list_with_specials_in_encoded_word(self):
2950+
# An encoded-word parsed from a structured header must remain
2951+
# encoded when it contains specials. Regression for gh-121284.
2952+
policy = self.policy.clone(max_line_length=40)
2953+
cases = [
2954+
# (to, folded)
2955+
('=?utf-8?q?A_v=C3=A9ry_long_name_with=2C_comma?= <to@example.com>',
2956+
'A =?utf-8?q?v=C3=A9ry_long_name_with?=\n'
2957+
' =?utf-8?q?=2C?= comma <to@example.com>\n'),
2958+
('=?utf-8?q?This_long_name_does_not_need_encoded=2Dword?= <to@example.com>',
2959+
'This long name does not need\n'
2960+
' encoded-word <to@example.com>\n'),
2961+
('"A véry long name with, comma" <to@example.com>',
2962+
# (This isn't the best fold point, but it's not invalid.)
2963+
'A =?utf-8?q?v=C3=A9ry_long_name_with?=\n'
2964+
' =?utf-8?q?=2C?= comma <to@example.com>\n'),
2965+
('"A véry long name containing a, comma" <to@example.com>',
2966+
'A =?utf-8?q?v=C3=A9ry?= long name\n'
2967+
' containing =?utf-8?q?a=2C?= comma\n'
2968+
' <to@example.com>\n'),
2969+
]
2970+
for (to, folded) in cases:
2971+
with self.subTest(to=to):
2972+
self._test(parser.get_address_list(to)[0], folded, policy=policy)
2973+
29492974
# XXX Need tests with comments on various sides of a unicode token,
29502975
# and with unicode tokens in the comments. Spaces inside the quotes
29512976
# currently don't do the right thing.

‎Misc/NEWS.d/next/Security/2024-08-06-12-27-34.gh-issue-121284.8rwPxe.rst

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,7 @@
1+
Fix bug in the folding of rfc2047 encoded-words when flattening an email message
2+
using a modern email policy. Previously when an encoded-word was too long
3+
for a line, it would be decoded, split across lines, and re-encoded. But commas
4+
and other special characters in the original text could be left unencoded and
5+
unquoted. This could theoretically be used to spoof header lines using
6+
a carefully constructed encoded-word if the resulting rendered email was
7+
transmitted or re-parsed.

0 commit comments

Comments
 (0)
0