|
 |
d61550 |
From 2d4a3fee6de2fa45eb82169361918f759269b4ec Mon Sep 17 00:00:00 2001
|
|
|
d61550 |
From: Seth Michael Larson <sethmichaellarson@gmail.com>
|
|
|
d61550 |
Date: Wed, 26 May 2021 10:43:12 -0500
|
|
|
d61550 |
Subject: [PATCH] Improve performance of sub-authority splitting in URL
|
|
|
d61550 |
|
|
|
d61550 |
---
|
|
|
d61550 |
src/urllib3/util/url.py | 8 +++++---
|
|
|
d61550 |
test/test_util.py | 10 ++++++++++
|
|
|
d61550 |
2 files changed, 15 insertions(+), 3 deletions(-)
|
|
|
d61550 |
|
|
|
d61550 |
diff --git a/src/urllib3/util/url.py b/src/urllib3/util/url.py
|
|
|
d61550 |
index 6ff238fe3c..81a03da9e3 100644
|
|
|
d61550 |
--- a/src/urllib3/util/url.py
|
|
|
d61550 |
+++ b/src/urllib3/util/url.py
|
|
|
d61550 |
@@ -63,12 +63,12 @@
|
|
|
d61550 |
BRACELESS_IPV6_ADDRZ_RE = re.compile("^" + IPV6_ADDRZ_PAT[2:-2] + "$")
|
|
|
d61550 |
ZONE_ID_RE = re.compile("(" + ZONE_ID_PAT + r")\]$")
|
|
|
d61550 |
|
|
|
d61550 |
-SUBAUTHORITY_PAT = (u"^(?:(.*)@)?(%s|%s|%s)(?::([0-9]{0,5}))?$") % (
|
|
|
d61550 |
+_HOST_PORT_PAT = ("^(%s|%s|%s)(?::([0-9]{0,5}))?$") % (
|
|
|
d61550 |
REG_NAME_PAT,
|
|
|
d61550 |
IPV4_PAT,
|
|
|
d61550 |
IPV6_ADDRZ_PAT,
|
|
|
d61550 |
)
|
|
|
d61550 |
-SUBAUTHORITY_RE = re.compile(SUBAUTHORITY_PAT, re.UNICODE | re.DOTALL)
|
|
|
d61550 |
+_HOST_PORT_RE = re.compile(_HOST_PORT_PAT, re.UNICODE | re.DOTALL)
|
|
|
d61550 |
|
|
|
d61550 |
UNRESERVED_CHARS = set(
|
|
|
d61550 |
"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789._-~"
|
|
|
d61550 |
@@ -365,7 +365,9 @@ def parse_url(url):
|
|
|
d61550 |
scheme = scheme.lower()
|
|
|
d61550 |
|
|
|
d61550 |
if authority:
|
|
|
d61550 |
- auth, host, port = SUBAUTHORITY_RE.match(authority).groups()
|
|
|
d61550 |
+ auth, _, host_port = authority.rpartition("@")
|
|
|
d61550 |
+ auth = auth or None
|
|
|
d61550 |
+ host, port = _HOST_PORT_RE.match(host_port).groups()
|
|
|
d61550 |
if auth and normalize_uri:
|
|
|
d61550 |
auth = _encode_invalid_chars(auth, USERINFO_CHARS)
|
|
|
d61550 |
if port == "":
|
|
|
d61550 |
diff --git a/test/test_util.py b/test/test_util.py
|
|
|
d61550 |
index a5b68a084b..88409e2d6c 100644
|
|
|
d61550 |
--- a/test/test_util.py
|
|
|
d61550 |
+++ b/test/test_util.py
|
|
|
d61550 |
@@ -438,6 +438,16 @@ def test_netloc(self, url, expected_netloc):
|
|
|
d61550 |
fragment="hash",
|
|
|
d61550 |
),
|
|
|
d61550 |
),
|
|
|
d61550 |
+ # Tons of '@' causing backtracking
|
|
|
d61550 |
+ ("https://" + ("@" * 10000) + "[", False),
|
|
|
d61550 |
+ (
|
|
|
d61550 |
+ "https://user:" + ("@" * 10000) + "example.com",
|
|
|
d61550 |
+ Url(
|
|
|
d61550 |
+ scheme="https",
|
|
|
d61550 |
+ auth="user:" + ("%40" * 9999),
|
|
|
d61550 |
+ host="example.com",
|
|
|
d61550 |
+ ),
|
|
|
d61550 |
+ ),
|
|
|
d61550 |
]
|
|
|
d61550 |
|
|
|
d61550 |
@pytest.mark.parametrize("url, expected_url", url_vulnerabilities)
|