bpo-30458: Disallow control chars in http URLs. (GH-12755) (GH-13154)
Disallow control chars in http URLs in urllib.urlopen. This addresses a potential security problem for applications that do not sanity check their URLs where http request headers could be injected.
Disable https related urllib tests on a build without ssl (GH-13032)
These tests require an SSL enabled build. Skip these tests when python is built without SSL to fix test failures.
Use http.client.InvalidURL instead of ValueError as the new error case's exception. (GH-13044)
Backport Co-Authored-By: Miro Hrončok <miro@hroncok.cz>
diff --git a/Lib/test/test_urllib.py b/Lib/test/test_urllib.py
index 2ac73b5..7214492 100644
--- a/Lib/test/test_urllib.py
+++ b/Lib/test/test_urllib.py
@@ -329,6 +329,59 @@
finally:
self.unfakehttp()
+ @unittest.skipUnless(ssl, "ssl module required")
+ def test_url_with_control_char_rejected(self):
+ for char_no in list(range(0, 0x21)) + [0x7f]:
+ char = chr(char_no)
+ schemeless_url = f"//localhost:7777/test{char}/"
+ self.fakehttp(b"HTTP/1.1 200 OK\r\n\r\nHello.")
+ try:
+ # We explicitly test urllib.request.urlopen() instead of the top
+ # level 'def urlopen()' function defined in this... (quite ugly)
+ # test suite. They use different url opening codepaths. Plain
+ # urlopen uses FancyURLOpener which goes via a codepath that
+ # calls urllib.parse.quote() on the URL which makes all of the
+ # above attempts at injection within the url _path_ safe.
+ escaped_char_repr = repr(char).replace('\\', r'\\')
+ InvalidURL = http.client.InvalidURL
+ with self.assertRaisesRegex(
+ InvalidURL, f"contain control.*{escaped_char_repr}"):
+ urllib.request.urlopen(f"http:{schemeless_url}")
+ with self.assertRaisesRegex(
+ InvalidURL, f"contain control.*{escaped_char_repr}"):
+ urllib.request.urlopen(f"https:{schemeless_url}")
+ # This code path quotes the URL so there is no injection.
+ resp = urlopen(f"http:{schemeless_url}")
+ self.assertNotIn(char, resp.geturl())
+ finally:
+ self.unfakehttp()
+
+ @unittest.skipUnless(ssl, "ssl module required")
+ def test_url_with_newline_header_injection_rejected(self):
+ self.fakehttp(b"HTTP/1.1 200 OK\r\n\r\nHello.")
+ host = "localhost:7777?a=1 HTTP/1.1\r\nX-injected: header\r\nTEST: 123"
+ schemeless_url = "//" + host + ":8080/test/?test=a"
+ try:
+ # We explicitly test urllib.request.urlopen() instead of the top
+ # level 'def urlopen()' function defined in this... (quite ugly)
+ # test suite. They use different url opening codepaths. Plain
+ # urlopen uses FancyURLOpener which goes via a codepath that
+ # calls urllib.parse.quote() on the URL which makes all of the
+ # above attempts at injection within the url _path_ safe.
+ InvalidURL = http.client.InvalidURL
+ with self.assertRaisesRegex(
+ InvalidURL, r"contain control.*\\r.*(found at least . .)"):
+ urllib.request.urlopen(f"http:{schemeless_url}")
+ with self.assertRaisesRegex(InvalidURL, r"contain control.*\\n"):
+ urllib.request.urlopen(f"https:{schemeless_url}")
+ # This code path quotes the URL so there is no injection.
+ resp = urlopen(f"http:{schemeless_url}")
+ self.assertNotIn(' ', resp.geturl())
+ self.assertNotIn('\r', resp.geturl())
+ self.assertNotIn('\n', resp.geturl())
+ finally:
+ self.unfakehttp()
+
def test_read_0_9(self):
# "0.9" response accepted (but not "simple responses" without
# a status line)