package/python: add upstream security fix for CVE-2019-9948
PATCH 1/3 package/python: add upstream security fix for CVE-2019-9948
Fixes CVE-2019-9948: Unnecessary URL scheme exists to allow file:// reading file in urllib.
https://bugs.python.org/issue35907
--- ...VE-2019-9948-urllib-rejects-local_file-sc.patch | 59 ++++++++++++++++++++++ 1 file changed, 59 insertions(+) create mode 100644 package/python/0035-bpo-35907-CVE-2019-9948-urllib-rejects-local_file-sc.patch
package/python/0035-bpo-35907-CVE-2019-9948-urllib-rejects-local_file-sc.patch new file +59 -0
@@ -0,0 +1,59 @@ +From b15bde8058e821b383d81fcae68b335a752083ca Mon Sep 17 00:00:00 2001 +From: SH <[email protected]> +Date: Wed, 22 May 2019 06:12:23 +0900 +Subject: [PATCH] bpo-35907, CVE-2019-9948: urllib rejects local_file:// scheme + (GH-11842) + + CVE-2019-9948: Avoid file reading as disallowing the unnecessary URL scheme in urllib.urlopen(). + +Signed-off-by: Peter Korsgaard <[email protected]> +--- + Lib/test/test_urllib.py | 7 +++++++ + Lib/urllib.py | 4 +++- + Misc/NEWS.d/next/Library/2019-02-13-17-21-10.bpo-35907.ckk2zg.rst | 1 + + 3 files changed, 11 insertions(+), 1 deletion(-) + create mode 100644 Misc/NEWS.d/next/Library/2019-02-13-17-21-10.bpo-35907.ckk2zg.rst + +diff --git a/Lib/test/test_urllib.py b/Lib/test/test_urllib.py +index d7778d4194..ae1f6c0b29 100644 +--- a/Lib/test/test_urllib.py ++++ b/Lib/test/test_urllib.py +@@ -1048,6 +1048,13 @@ class URLopener_Tests(unittest.TestCase): + "spam://c:|windows%/:=&?~#+!$,;'@()*[]|/path/"), + "//c:|windows%/:=&?~#+!$,;'@()*[]|/path/") + ++ def test_local_file_open(self): ++ class DummyURLopener(urllib.URLopener): ++ def open_local_file(self, url): ++ return url ++ for url in ('local_file://example', 'local-file://example'): ++ self.assertRaises(IOError, DummyURLopener().open, url) ++ self.assertRaises(IOError, urllib.urlopen, url) + + # Just commented them out. + # Can't really tell why keep failing in windows and sparc. +diff --git a/Lib/urllib.py b/Lib/urllib.py +index d85504a5cb..156879dd0a 100644 +--- a/Lib/urllib.py ++++ b/Lib/urllib.py +@@ -203,7 +203,9 @@ class URLopener: + name = 'open_' + urltype + self.type = urltype + name = name.replace('-', '_') +- if not hasattr(self, name): ++ ++ # bpo-35907: disallow the file reading with the type not allowed ++ if not hasattr(self, name) or name == 'open_local_file': + if proxy: + return self.open_unknown_proxy(proxy, fullurl, data) + else: +diff --git a/Misc/NEWS.d/next/Library/2019-02-13-17-21-10.bpo-35907.ckk2zg.rst b/Misc/NEWS.d/next/Library/2019-02-13-17-21-10.bpo-35907.ckk2zg.rst +new file mode 100644 +index 0000000000..bb187d8d65 +--- /dev/null ++++ b/Misc/NEWS.d/next/Library/2019-02-13-17-21-10.bpo-35907.ckk2zg.rst +@@ -0,0 +1 @@ ++CVE-2019-9948: Avoid file reading as disallowing the unnecessary URL scheme in urllib.urlopen +-- +2.11.0 +
Signature · 1 line
2.11.0
PATCH 1/3 package/python: add upstream security fix for CVE-2019-9948
Quoted 8 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Peter" == Peter Korsgaard <[email protected]> writes:
Fixes CVE-2019-9948: Unnecessary URL scheme exists to allow file:// reading file in urllib.
https://bugs.python.org/issue35907
Signed-off-by Peter Korsgaard <[email protected]>
Committed to 2019.02.x and 2019.05.x, thanks.
Signature · 1 line
Bye, Peter Korsgaard
PATCH 2/3 package/python: add upstream security fix for CVE-2019-9636
In reply to Peter KorsgaardFixes CVE-2019-9636: urlsplit does not handle NFKC normalization
https://bugs.python.org/issue36216
The fix unfortunately introduced regressions, so also apply the followup fixes.
https://bugs.python.org/issue36742
--- ...dd-check-for-characters-in-netloc-that-no.patch | 159 +++++++++++++++++++++ ...16-Only-print-test-messages-when-verbose-.patch | 28 ++++ ...ixes-handling-of-pre-normalization-charac.patch | 66 +++++++++ ...orrects-fix-to-handle-decomposition-in-us.patch | 67 +++++++++ ...42-Fix-urlparse.urlsplit-error-message-fo.patch | 67 +++++++++ 5 files changed, 387 insertions(+) create mode 100644 package/python/0036-bpo-36216-Add-check-for-characters-in-netloc-that-no.patch create mode 100644 package/python/0037-3.7-bpo-36216-Only-print-test-messages-when-verbose-.patch create mode 100644 package/python/0038-bpo-36742-Fixes-handling-of-pre-normalization-charac.patch create mode 100644 package/python/0039-bpo-36742-Corrects-fix-to-handle-decomposition-in-us.patch create mode 100644 package/python/0040-2.7-bpo-36742-Fix-urlparse.urlsplit-error-message-fo.patch
package/python/0036-bpo-36216-Add-check-for-characters-in-netloc-that-no.patch new file +159 -0
@@ -0,0 +1,159 @@ +From e37ef41289b77e0f0bb9a6aedb0360664c55bdd5 Mon Sep 17 00:00:00 2001 +From: Steve Dower <[email protected]> +Date: Thu, 7 Mar 2019 09:08:45 -0800 +Subject: [PATCH] bpo-36216: Add check for characters in netloc that normalize + to separators (GH-12201) + +Signed-off-by: Peter Korsgaard <[email protected]> +--- + Doc/library/urlparse.rst | 20 ++++++++++++++++++ + Lib/test/test_urlparse.py | 24 ++++++++++++++++++++++ + Lib/urlparse.py | 17 +++++++++++++++ + .../2019-03-06-09-38-40.bpo-36216.6q1m4a.rst | 3 +++ + 4 files changed, 64 insertions(+) + create mode 100644 Misc/NEWS.d/next/Security/2019-03-06-09-38-40.bpo-36216.6q1m4a.rst + +diff --git a/Doc/library/urlparse.rst b/Doc/library/urlparse.rst +index 22249da54f..0989c88c30 100644 +--- a/Doc/library/urlparse.rst ++++ b/Doc/library/urlparse.rst +@@ -119,12 +119,22 @@ The :mod:`urlparse` module defines the following functions: + See section :ref:`urlparse-result-object` for more information on the result + object. + ++ Characters in the :attr:`netloc` attribute that decompose under NFKC ++ normalization (as used by the IDNA encoding) into any of ``/``, ``?``, ++ ``#``, ``@``, or ``:`` will raise a :exc:`ValueError`. If the URL is ++ decomposed before parsing, or is not a Unicode string, no error will be ++ raised. ++ + .. versionchanged:: 2.5 + Added attributes to return value. + + .. versionchanged:: 2.7 + Added IPv6 URL parsing capabilities. + ++ .. versionchanged:: 2.7.17 ++ Characters that affect netloc parsing under NFKC normalization will ++ now raise :exc:`ValueError`. ++ + + .. function:: parse_qs(qs[, keep_blank_values[, strict_parsing[, max_num_fields]]]) + +@@ -232,11 +242,21 @@ The :mod:`urlparse` module defines the following functions: + See section :ref:`urlparse-result-object` for more information on the result + object. + ++ Characters in the :attr:`netloc` attribute that decompose under NFKC ++ normalization (as used by the IDNA encoding) into any of ``/``, ``?``, ++ ``#``, ``@``, or ``:`` will raise a :exc:`ValueError`. If the URL is ++ decomposed before parsing, or is not a Unicode string, no error will be ++ raised. ++ + .. versionadded:: 2.2 + + .. versionchanged:: 2.5 + Added attributes to return value. + ++ .. versionchanged:: 2.7.17 ++ Characters that affect netloc parsing under NFKC normalization will ++ now raise :exc:`ValueError`. ++ + + .. function:: urlunsplit(parts) + +diff --git a/Lib/test/test_urlparse.py b/Lib/test/test_urlparse.py +index 4e1ded73c2..73b0228ea8 100644 +--- a/Lib/test/test_urlparse.py ++++ b/Lib/test/test_urlparse.py +@@ -1,4 +1,6 @@ + from test import test_support ++import sys ++import unicodedata + import unittest + import urlparse + +@@ -624,6 +626,28 @@ class UrlParseTestCase(unittest.TestCase): + self.assertEqual(urlparse.urlparse("http://www.python.org:80"), + ('http','www.python.org:80','','','','')) + ++ def test_urlsplit_normalization(self): ++ # Certain characters should never occur in the netloc, ++ # including under normalization. ++ # Ensure that ALL of them are detected and cause an error ++ illegal_chars = u'/:#?@' ++ hex_chars = {'{:04X}'.format(ord(c)) for c in illegal_chars} ++ denorm_chars = [ ++ c for c in map(unichr, range(128, sys.maxunicode)) ++ if (hex_chars & set(unicodedata.decomposition(c).split())) ++ and c not in illegal_chars ++ ] ++ # Sanity check that we found at least one such character ++ self.assertIn(u'\u2100', denorm_chars) ++ self.assertIn(u'\uFF03', denorm_chars) ++ ++ for scheme in [u"http", u"https", u"ftp"]: ++ for c in denorm_chars: ++ url = u"{}://netloc{}false.netloc/path".format(scheme, c) ++ print "Checking %r" % url ++ with self.assertRaises(ValueError): ++ urlparse.urlsplit(url) ++ + def test_main(): + test_support.run_unittest(UrlParseTestCase) + +diff --git a/Lib/urlparse.py b/Lib/urlparse.py +index f7c2b032b0..54eda08651 100644 +--- a/Lib/urlparse.py ++++ b/Lib/urlparse.py +@@ -165,6 +165,21 @@ def _splitnetloc(url, start=0): + delim = min(delim, wdelim) # use earliest delim position + return url[start:delim], url[delim:] # return (domain, rest) + ++def _checknetloc(netloc): ++ if not netloc or not isinstance(netloc, unicode): ++ return ++ # looking for characters like \u2100 that expand to 'a/c' ++ # IDNA uses NFKC equivalence, so normalize for this check ++ import unicodedata ++ netloc2 = unicodedata.normalize('NFKC', netloc) ++ if netloc == netloc2: ++ return ++ _, _, netloc = netloc.rpartition('@') # anything to the left of '@' is okay ++ for c in '/?#@:': ++ if c in netloc2: ++ raise ValueError("netloc '" + netloc2 + "' contains invalid " + ++ "characters under NFKC normalization") ++ + def urlsplit(url, scheme='', allow_fragments=True): + """Parse a URL into 5 components: + <scheme>://<netloc>/<path>?<query>#<fragment> +@@ -193,6 +208,7 @@ def urlsplit(url, scheme='', allow_fragments=True): + url, fragment = url.split('#', 1) + if '?' in url: + url, query = url.split('?', 1) ++ _checknetloc(netloc) + v = SplitResult(scheme, netloc, url, query, fragment) + _parse_cache[key] = v + return v +@@ -216,6 +232,7 @@ def urlsplit(url, scheme='', allow_fragments=True): + url, fragment = url.split('#', 1) + if '?' in url: + url, query = url.split('?', 1) ++ _checknetloc(netloc) + v = SplitResult(scheme, netloc, url, query, fragment) + _parse_cache[key] = v + return v +diff --git a/Misc/NEWS.d/next/Security/2019-03-06-09-38-40.bpo-36216.6q1m4a.rst b/Misc/NEWS.d/next/Security/2019-03-06-09-38-40.bpo-36216.6q1m4a.rst +new file mode 100644 +index 0000000000..1e1ad92c6f +--- /dev/null ++++ b/Misc/NEWS.d/next/Security/2019-03-06-09-38-40.bpo-36216.6q1m4a.rst +@@ -0,0 +1,3 @@ ++Changes urlsplit() to raise ValueError when the URL contains characters that ++decompose under IDNA encoding (NFKC-normalization) into characters that ++affect how the URL is parsed. +\ No newline at end of file +-- +2.11.0 +
package/python/0037-3.7-bpo-36216-Only-print-test-messages-when-verbose-.patch new file +28 -0
@@ -0,0 +1,28 @@ +From 507bd8cde60ced74d13a1ffa883bb9b0e73c38be Mon Sep 17 00:00:00 2001 +From: Steve Dower <[email protected]> +Date: Tue, 12 Mar 2019 13:51:58 -0700 +Subject: [PATCH] [3.7] bpo-36216: Only print test messages when verbose + (GH-12291) + +Signed-off-by: Peter Korsgaard <[email protected]> +--- + Lib/test/test_urlparse.py | 3 ++- + 1 file changed, 2 insertions(+), 1 deletion(-) + +diff --git a/Lib/test/test_urlparse.py b/Lib/test/test_urlparse.py +index 73b0228ea8..1830d0b286 100644 +--- a/Lib/test/test_urlparse.py ++++ b/Lib/test/test_urlparse.py +@@ -644,7 +644,8 @@ class UrlParseTestCase(unittest.TestCase): + for scheme in [u"http", u"https", u"ftp"]: + for c in denorm_chars: + url = u"{}://netloc{}false.netloc/path".format(scheme, c) +- print "Checking %r" % url ++ if test_support.verbose: ++ print "Checking %r" % url + with self.assertRaises(ValueError): + urlparse.urlsplit(url) + +-- +2.11.0 +
package/python/0038-bpo-36742-Fixes-handling-of-pre-normalization-charac.patch new file +66 -0
66 lines not on this page · raw message
package/python/0039-bpo-36742-Corrects-fix-to-handle-decomposition-in-us.patch new file +67 -0
67 lines not on this page · raw message
package/python/0040-2.7-bpo-36742-Fix-urlparse.urlsplit-error-message-fo.patch new file +67 -0
67 lines not on this page · raw message
Signature · 1 line
2.11.0
PATCH 2/3 package/python: add upstream security fix for CVE-2019-9636
Quoted 11 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Peter" == Peter Korsgaard <[email protected]> writes:
Fixes CVE-2019-9636: urlsplit does not handle NFKC normalization
https://bugs.python.org/issue36216The fix unfortunately introduced regressions, so also apply the followup fixes.
https://bugs.python.org/issue36742
Signed-off-by Peter Korsgaard <[email protected]>
Committed to 2019.02.x and 2019.05.x, thanks.
Signature · 1 line
Bye, Peter Korsgaard
PATCH 1/3 package/python: add upstream security fix for CVE-2019-9948
In reply to Peter KorsgaardQuoted 10 lines · On Sun, 16 Jun 2019 23:17:09 +0200 Peter Korsgaard <[email protected]> wrote:
Fixes CVE-2019-9948: Unnecessary URL scheme exists to allow file:// reading file in urllib.
https://bugs.python.org/issue35907
Signed-off-by Peter Korsgaard <[email protected]>--- ...VE-2019-9948-urllib-rejects-local_file-sc.patch | 59 ++++++++++++++++++++++ 1 file changed, 59 insertions(+) create mode 100644 package/python/0035-bpo-35907-CVE-2019-9948-urllib-rejects-local_file-sc.patch
All three patches applied to master. Thanks!
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
PATCH 3/3 package/python3: add upstream security fix for CVE-2019-10160
In reply to Peter KorsgaardFixes CVE-2019-10160: urlsplit does not handle NFKC normalization (2nd fix)
While the fix for CVE-2019-9936 is included in 3.7.3, the followup regression fixes unfortunatly aren't.
https://bugs.python.org/issue36742
--- ...ixes-handling-of-pre-normalization-charac.patch | 70 ++++++++++++++++++++++ ...orrects-fix-to-handle-decomposition-in-us.patch | 58 ++++++++++++++++++ 2 files changed, 128 insertions(+) create mode 100644 package/python3/0033-bpo-36742-Fixes-handling-of-pre-normalization-charac.patch create mode 100644 package/python3/0034-bpo-36742-Corrects-fix-to-handle-decomposition-in-us.patch
package/python3/0033-bpo-36742-Fixes-handling-of-pre-normalization-charac.patch new file +70 -0
@@ -0,0 +1,70 @@ +From 4d723e76e1ad17e9e7d5e828e59bb47e76f2174b Mon Sep 17 00:00:00 2001 +From: "Miss Islington (bot)" + <[email protected]> +Date: Tue, 30 Apr 2019 05:21:02 -0700 +Subject: [PATCH] bpo-36742: Fixes handling of pre-normalization characters in + urlsplit() (GH-13017) + +(cherry picked from commit d537ab0ff9767ef024f26246899728f0116b1ec3) + +Co-authored-by: Steve Dower <[email protected]> +Signed-off-by: Peter Korsgaard <[email protected]> +--- + Lib/test/test_urlparse.py | 6 ++++++ + Lib/urllib/parse.py | 11 +++++++---- + .../next/Security/2019-04-29-15-34-59.bpo-36742.QCUY0i.rst | 1 + + 3 files changed, 14 insertions(+), 4 deletions(-) + create mode 100644 Misc/NEWS.d/next/Security/2019-04-29-15-34-59.bpo-36742.QCUY0i.rst + +diff --git a/Lib/test/test_urlparse.py b/Lib/test/test_urlparse.py +index e6638aee22..c262354494 100644 +--- a/Lib/test/test_urlparse.py ++++ b/Lib/test/test_urlparse.py +@@ -1001,6 +1001,12 @@ class UrlParseTestCase(unittest.TestCase): + self.assertIn('\u2100', denorm_chars) + self.assertIn('\uFF03', denorm_chars) + ++ # bpo-36742: Verify port separators are ignored when they ++ # existed prior to decomposition ++ urllib.parse.urlsplit('http://\u30d5\u309a:80') ++ with self.assertRaises(ValueError): ++ urllib.parse.urlsplit('http://\u30d5\u309a\ufe1380') ++ + for scheme in ["http", "https", "ftp"]: + for c in denorm_chars: + url = "{}://netloc{}false.netloc/path".format(scheme, c) +diff --git a/Lib/urllib/parse.py b/Lib/urllib/parse.py +index 1eec26e0f1..f5b3487ea9 100644 +--- a/Lib/urllib/parse.py ++++ b/Lib/urllib/parse.py +@@ -397,13 +397,16 @@ def _checknetloc(netloc): + # looking for characters like \u2100 that expand to 'a/c' + # IDNA uses NFKC equivalence, so normalize for this check + import unicodedata +- netloc2 = unicodedata.normalize('NFKC', netloc) +- if netloc == netloc2: ++ n = netloc.rpartition('@')[2] # ignore anything to the left of '@' ++ n = n.replace(':', '') # ignore characters already included ++ n = n.replace('#', '') # but not the surrounding text ++ n = n.replace('?', '') ++ netloc2 = unicodedata.normalize('NFKC', n) ++ if n == netloc2: + return +- _, _, netloc = netloc.rpartition('@') # anything to the left of '@' is okay + for c in '/?#@:': + if c in netloc2: +- raise ValueError("netloc '" + netloc2 + "' contains invalid " + ++ raise ValueError("netloc '" + netloc + "' contains invalid " + + "characters under NFKC normalization") + + def urlsplit(url, scheme='', allow_fragments=True): +diff --git a/Misc/NEWS.d/next/Security/2019-04-29-15-34-59.bpo-36742.QCUY0i.rst b/Misc/NEWS.d/next/Security/2019-04-29-15-34-59.bpo-36742.QCUY0i.rst +new file mode 100644 +index 0000000000..d729ed2f3c +--- /dev/null ++++ b/Misc/NEWS.d/next/Security/2019-04-29-15-34-59.bpo-36742.QCUY0i.rst +@@ -0,0 +1 @@ ++Fixes mishandling of pre-normalization characters in urlsplit(). +-- +2.11.0 +
package/python3/0034-bpo-36742-Corrects-fix-to-handle-decomposition-in-us.patch new file +58 -0
58 lines not on this page · raw message
Signature · 1 line
2.11.0
PATCH 3/3 package/python3: add upstream security fix for CVE-2019-10160
Quoted 9 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Peter" == Peter Korsgaard <[email protected]> writes:
Fixes CVE-2019-10160: urlsplit does not handle NFKC normalization (2nd fix) While the fix for CVE-2019-9936 is included in 3.7.3, the followup regression fixes unfortunatly aren't.
https://bugs.python.org/issue36742
Signed-off-by Peter Korsgaard <[email protected]>
Committed to 2019.02.x and 2019.05.x, thanks.
Signature · 1 line
Bye, Peter Korsgaard