[Python-checkins] gh-100001: Omit control characters in http.server stderr logs. (#100002)

Mon Dec 5 15:55:51 EST 2022

https://github.com/python/cpython/commit/d8ab0a4dfa48f881b4ac9ab857d2e9de42f72828
commit: d8ab0a4dfa48f881b4ac9ab857d2e9de42f72828
branch: main
author: Gregory P. Smith <greg at krypto.org>
committer: gpshead <greg at krypto.org>
date: 2022年12月05日T12:55:45-08:00
summary:
gh-100001: Omit control characters in http.server stderr logs. (#100002)
Replace control characters in http.server.BaseHTTPRequestHandler.log_message with an escaped \xHH sequence to avoid causing problems for the terminal the output is printed to.
files:
A Misc/NEWS.d/next/Security/2022-12-05-01-39-10.gh-issue-100001.uD05Fc.rst
M Doc/library/http.server.rst
M Lib/http/server.py
M Lib/test/test_httpservers.py

diff --git a/Doc/library/http.server.rst b/Doc/library/http.server.rst
index 81b6bf5373b4..154f3f3c3904 100644
--- a/Doc/library/http.server.rst
+++ b/Doc/library/http.server.rst
@@ -512,3 +512,10 @@ Security Considerations
 :class:`SimpleHTTPRequestHandler` will follow symbolic links when handling
 requests, this makes it possible for files outside of the specified directory
 to be served.
+
+Earlier versions of Python did not scrub control characters from the
+log messages emitted to stderr from ``python -m http.server`` or the
+default :class:`BaseHTTPRequestHandler` ``.log_message``
+implementation. This could allow to remote clients connecting to your
+server to send nefarious control codes to your terminal.
+
diff --git a/Lib/http/server.py b/Lib/http/server.py
index 8aee31bac275..3b5bd9eb0129 100644
--- a/Lib/http/server.py
+++ b/Lib/http/server.py
@@ -93,6 +93,7 @@
 import html
 import http.client
 import io
+import itertools
 import mimetypes
 import os
 import posixpath
@@ -562,6 +563,10 @@ def log_error(self, format, *args):
 
 self.log_message(format, *args)
 
+ # https://en.wikipedia.org/wiki/List_of_Unicode_characters#Control_codes
+ _control_char_table = str.maketrans(
+ {c: fr'\x{c:02x}' for c in itertools.chain(range(0x20), range(0x7f,0xa0))})
+
 def log_message(self, format, *args):
 """Log an arbitrary message.
 
@@ -577,12 +582,16 @@ def log_message(self, format, *args):
 The client ip and current date/time are prefixed to
 every message.
 
+ Unicode control characters are replaced with escaped hex
+ before writing the output to stderr.
+
 """
 
+ message = format % args
 sys.stderr.write("%s - - [%s] %s\n" %
 (self.address_string(),
 self.log_date_time_string(),
- format%args))
+ message.translate(self._control_char_table)))
 
 def version_string(self):
 """Return the server software version string."""
diff --git a/Lib/test/test_httpservers.py b/Lib/test/test_httpservers.py
index a937258069ed..b0e2d713e3de 100644
--- a/Lib/test/test_httpservers.py
+++ b/Lib/test/test_httpservers.py
@@ -26,7 +26,7 @@
 import datetime
 import threading
 from unittest import mock
-from io import BytesIO
+from io import BytesIO, StringIO
 
 import unittest
 from test import support
@@ -990,6 +990,25 @@ def verify_http_server_response(self, response):
 match = self.HTTPResponseMatch.search(response)
 self.assertIsNotNone(match)
 
+ def test_unprintable_not_logged(self):
+ # We call the method from the class directly as our Socketless
+ # Handler subclass overrode it... nice for everything BUT this test.
+ self.handler.client_address = ('127.0.0.1', 1337)
+ log_message = BaseHTTPRequestHandler.log_message
+ with mock.patch.object(sys, 'stderr', StringIO()) as fake_stderr:
+ log_message(self.handler, '/foo')
+ log_message(self.handler, '/033円bar000円033円')
+ log_message(self.handler, '/spam %s.', 'a')
+ log_message(self.handler, '/spam %s.', '033円\x7f\x9f\xa0beans')
+ stderr = fake_stderr.getvalue()
+ self.assertNotIn('033円', stderr) # non-printable chars are caught.
+ self.assertNotIn('000円', stderr) # non-printable chars are caught.
+ lines = stderr.splitlines()
+ self.assertIn('/foo', lines[0])
+ self.assertIn(r'/\x1bbar\x00\x1b', lines[1])
+ self.assertIn('/spam a.', lines[2])
+ self.assertIn('/spam \\x1b\\x7f\\x9f\xa0beans.', lines[3])
+
 def test_http_1_1(self):
 result = self.send_typical_request(b'GET / HTTP/1.1\r\n\r\n')
 self.verify_http_server_response(result[0])
diff --git a/Misc/NEWS.d/next/Security/2022-12-05-01-39-10.gh-issue-100001.uD05Fc.rst b/Misc/NEWS.d/next/Security/2022-12-05-01-39-10.gh-issue-100001.uD05Fc.rst
new file mode 100644
index 000000000000..a396e95cd83f
--- /dev/null
+++ b/Misc/NEWS.d/next/Security/2022-12-05-01-39-10.gh-issue-100001.uD05Fc.rst
@@ -0,0 +1,6 @@
+``python -m http.server`` no longer allows terminal control characters sent
+within a garbage request to be printed to the stderr server log.
+
+This is done by changing the :mod:`http.server` :class:`BaseHTTPRequestHandler`
+``.log_message`` method to replace control characters with a ``\xHH`` hex escape
+before printing.
