Skip to content

Commit 526617e

Browse files
miss-islingtonsethmlarsonZeroIntensity
authored
[3.11] pythongh-105704: Disallow square brackets ([ and ]) in domain names for parsed URLs (pythonGH-129418) (python#129528)
Co-authored-by: Seth Michael Larson <[email protected]> Co-authored-by: Peter Bierma <[email protected]>
1 parent 7cff053 commit 526617e

File tree

3 files changed

+58
-3
lines changed

3 files changed

+58
-3
lines changed

‎Lib/test/test_urlparse.py‎

Lines changed: 36 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1224,16 +1224,51 @@ def test_invalid_bracketed_hosts(self):
12241224
self.assertRaises(ValueError, urllib.parse.urlsplit, 'Scheme://user@[0439:23af::2309::fae7:1234]/Path?Query')
12251225
self.assertRaises(ValueError, urllib.parse.urlsplit, 'Scheme://user@[0439:23af:2309::fae7:1234:2342:438e:192.0.2.146]/Path?Query')
12261226
self.assertRaises(ValueError, urllib.parse.urlsplit, 'Scheme://user@]v6a.ip[/Path')
1227+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[v6a.ip]')
1228+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[v6a.ip].suffix')
1229+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[v6a.ip]/')
1230+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[v6a.ip].suffix/')
1231+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[v6a.ip]?')
1232+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[v6a.ip].suffix?')
1233+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[::1]')
1234+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[::1].suffix')
1235+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[::1]/')
1236+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[::1].suffix/')
1237+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[::1]?')
1238+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[::1].suffix?')
1239+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[::1]:a')
1240+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[::1].suffix:a')
1241+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[::1]:a1')
1242+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[::1].suffix:a1')
1243+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[::1]:1a')
1244+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[::1].suffix:1a')
1245+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[::1]:')
1246+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[::1].suffix:/')
1247+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[::1]:?')
1248+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://user@prefix.[v6a.ip]')
1249+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://user@[v6a.ip].suffix')
1250+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://[v6a.ip')
1251+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://v6a.ip]')
1252+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://]v6a.ip[')
1253+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://]v6a.ip')
1254+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://v6a.ip[')
1255+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix.[v6a.ip')
1256+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://v6a.ip].suffix')
1257+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix]v6a.ip[suffix')
1258+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://prefix]v6a.ip')
1259+
self.assertRaises(ValueError, urllib.parse.urlsplit, 'scheme://v6a.ip[suffix')
12271260

12281261
def test_splitting_bracketed_hosts(self):
1229-
p1 = urllib.parse.urlsplit('scheme://user@[v6a.ip]/path?query')
1262+
p1 = urllib.parse.urlsplit('scheme://user@[v6a.ip]:1234/path?query')
12301263
self.assertEqual(p1.hostname, 'v6a.ip')
12311264
self.assertEqual(p1.username, 'user')
12321265
self.assertEqual(p1.path, '/path')
1266+
self.assertEqual(p1.port, 1234)
12331267
p2 = urllib.parse.urlsplit('scheme://user@[0439:23af:2309::fae7%test]/path?query')
12341268
self.assertEqual(p2.hostname, '0439:23af:2309::fae7%test')
12351269
self.assertEqual(p2.username, 'user')
12361270
self.assertEqual(p2.path, '/path')
1271+
self.assertIs(p2.port, None)
12371272
p3 = urllib.parse.urlsplit('scheme://user@[0439:23af:2309::fae7:1234:192.0.2.146%test]/path?query')
12381273
self.assertEqual(p3.hostname, '0439:23af:2309::fae7:1234:192.0.2.146%test')
12391274
self.assertEqual(p3.username, 'user')

‎Lib/urllib/parse.py‎

Lines changed: 18 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -436,6 +436,23 @@ def _checknetloc(netloc):
436436
raise ValueError("netloc '" + netloc + "' contains invalid " +
437437
"characters under NFKC normalization")
438438

439+
def _check_bracketed_netloc(netloc):
440+
# Note that this function must mirror the splitting
441+
# done in NetlocResultMixins._hostinfo().
442+
hostname_and_port = netloc.rpartition('@')[2]
443+
before_bracket, have_open_br, bracketed = hostname_and_port.partition('[')
444+
if have_open_br:
445+
# No data is allowed before a bracket.
446+
if before_bracket:
447+
raise ValueError("Invalid IPv6 URL")
448+
hostname, _, port = bracketed.partition(']')
449+
# No data is allowed after the bracket but before the port delimiter.
450+
if port and not port.startswith(":"):
451+
raise ValueError("Invalid IPv6 URL")
452+
else:
453+
hostname, _, port = hostname_and_port.partition(':')
454+
_check_bracketed_host(hostname)
455+
439456
# Valid bracketed hosts are defined in
440457
# https://www.rfc-editor.org/rfc/rfc3986#page-49 and https://url.spec.whatwg.org/
441458
def _check_bracketed_host(hostname):
@@ -496,8 +513,7 @@ def urlsplit(url, scheme='', allow_fragments=True):
496513
(']' in netloc and '[' not in netloc)):
497514
raise ValueError("Invalid IPv6 URL")
498515
if '[' in netloc and ']' in netloc:
499-
bracketed_host = netloc.partition('[')[2].partition(']')[0]
500-
_check_bracketed_host(bracketed_host)
516+
_check_bracketed_netloc(netloc)
501517
if allow_fragments and '#' in url:
502518
url, fragment = url.split('#', 1)
503519
if '?' in url:

‎Misc/NEWS.d/next/Security/2025-01-28-14-08-03.gh-issue-105704.EnhHxu.rst‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,4 @@
1+
When using :func:`urllib.parse.urlsplit` and :func:`urllib.parse.urlparse` host
2+
parsing would not reject domain names containing square brackets (``[`` and
3+
``]``). Square brackets are only valid for IPv6 and IPvFuture hosts according to
4+
`RFC 3986 Section 3.2.2 <https://www.rfc-editor.org/rfc/rfc3986#section-3.2.2>`__.

0 commit comments

Comments
 (0)