bpo-38576: Disallow control characters in hostnames in http.client (GH-18995) (GH-19002)

Add host validation for control characters for more CVE-2019-18348 protection.
(cherry picked from commit 9165addc22d05e776a54319a8531ebd0b2fe01ef)

Co-authored-by: Ashwin Ramaswami <aramaswamis@gmail.com>

2 files changed, 46 insertions, 3 deletions

diff --git a/Lib/test/test_httplib.py b/Lib/test/test_httplib.py
index 14d42d4837..fcd9231666 100644
--- a/Lib/test/test_httplib.py
+++ b/Lib/test/test_httplib.py
@@ -1132,7 +1132,7 @@ class BasicTest(TestCase):
         thread.join()
         self.assertEqual(result, b"proxied data\n")
 
-    def test_putrequest_override_validation(self):
+    def test_putrequest_override_domain_validation(self):
         """
         It should be possible to override the default validation
         behavior in putrequest (bpo-38216).
@@ -1145,6 +1145,17 @@ class BasicTest(TestCase):
         conn.sock = FakeSocket('')
         conn.putrequest('GET', '/\x00')
 
+    def test_putrequest_override_host_validation(self):
+        class UnsafeHTTPConnection(client.HTTPConnection):
+            def _validate_host(self, url):
+                pass
+
+        conn = UnsafeHTTPConnection('example.com\r\n')
+        conn.sock = FakeSocket('')
+        # set skip_host so a ValueError is not raised upon adding the
+        # invalid URL as the value of the "Host:" header
+        conn.putrequest('GET', '/', skip_host=1)
+
     def test_putrequest_override_encoding(self):
         """
         It should be possible to override the default encoding
diff --git a/Lib/test/test_urllib.py b/Lib/test/test_urllib.py
index 0061a5297c..ddf425fd8d 100644
--- a/Lib/test/test_urllib.py
+++ b/Lib/test/test_urllib.py
@@ -331,7 +331,7 @@ class urlopen_HttpTests(unittest.TestCase, FakeHTTPMixin, FakeFTPMixin):
             self.unfakehttp()
 
     @unittest.skipUnless(ssl, "ssl module required")
-    def test_url_with_control_char_rejected(self):
+    def test_url_path_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}/"
@@ -358,7 +358,7 @@ class urlopen_HttpTests(unittest.TestCase, FakeHTTPMixin, FakeFTPMixin):
                 self.unfakehttp()
 
     @unittest.skipUnless(ssl, "ssl module required")
-    def test_url_with_newline_header_injection_rejected(self):
+    def test_url_path_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"
@@ -383,6 +383,38 @@ class urlopen_HttpTests(unittest.TestCase, FakeHTTPMixin, FakeFTPMixin):
         finally:
             self.unfakehttp()
 
+    @unittest.skipUnless(ssl, "ssl module required")
+    def test_url_host_with_control_char_rejected(self):
+        for char_no in list(range(0, 0x21)) + [0x7f]:
+            char = chr(char_no)
+            schemeless_url = f"//localhost{char}/test/"
+            self.fakehttp(b"HTTP/1.1 200 OK\r\n\r\nHello.")
+            try:
+                escaped_char_repr = repr(char).replace('\\', r'\\')
+                InvalidURL = http.client.InvalidURL
+                with self.assertRaisesRegex(
+                    InvalidURL, f"contain control.*{escaped_char_repr}"):
+                    urlopen(f"http:{schemeless_url}")
+                with self.assertRaisesRegex(InvalidURL, f"contain control.*{escaped_char_repr}"):
+                    urlopen(f"https:{schemeless_url}")
+            finally:
+                self.unfakehttp()
+
+    @unittest.skipUnless(ssl, "ssl module required")
+    def test_url_host_with_newline_header_injection_rejected(self):
+        self.fakehttp(b"HTTP/1.1 200 OK\r\n\r\nHello.")
+        host = "localhost\r\nX-injected: header\r\n"
+        schemeless_url = "//" + host + ":8080/test/?test=a"
+        try:
+            InvalidURL = http.client.InvalidURL
+            with self.assertRaisesRegex(
+                InvalidURL, r"contain control.*\\r"):
+                urlopen(f"http:{schemeless_url}")
+            with self.assertRaisesRegex(InvalidURL, r"contain control.*\\n"):
+                urlopen(f"https:{schemeless_url}")
+        finally:
+            self.unfakehttp()
+
     def test_read_0_9(self):
         # "0.9" response accepted (but not "simple responses" without
         # a status line)