aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorMiss Islington (bot) <31488909+miss-islington@users.noreply.github.com>2022-11-13 11:00:25 -0800
committerMichał Górny <mgorny@gentoo.org>2024-01-15 14:39:10 +0100
commit715cd9e32f1f6ed2ed1dd17b5f89aa46495efc07 (patch)
tree24d50b705e23809089b3750f80075931e467ef68
parentgh-79096: Protect cookie file created by {LWP,Mozilla}CookieJar.save() (GH-93... (diff)
downloadpypy-715cd9e32f1f6ed2ed1dd17b5f89aa46495efc07.tar.gz
pypy-715cd9e32f1f6ed2ed1dd17b5f89aa46495efc07.tar.bz2
pypy-715cd9e32f1f6ed2ed1dd17b5f89aa46495efc07.zip
gh-99418: Make urllib.parse.urlparse enforce that a scheme must begin with an alphabetical ASCII character. (GH-99421)gentoo-3.9-7.3.15
Prevent urllib.parse.urlparse from accepting schemes that don't begin with an alphabetical ASCII character. RFC 3986 defines a scheme like this: `scheme = ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )` RFC 2234 defines an ALPHA like this: `ALPHA = %x41-5A / %x61-7A` The WHATWG URL spec defines a scheme like this: `"A URL-scheme string must be one ASCII alpha, followed by zero or more of ASCII alphanumeric, U+002B (+), U+002D (-), and U+002E (.)."` (cherry picked from commit 439b9cfaf43080e91c4ad69f312f21fa098befc7) Co-authored-by: Ben Kallus <49924171+kenballus@users.noreply.github.com>
-rw-r--r--lib-python/3/test/test_urlparse.py18
-rw-r--r--lib-python/3/urllib/parse.py2
2 files changed, 19 insertions, 1 deletions
diff --git a/lib-python/3/test/test_urlparse.py b/lib-python/3/test/test_urlparse.py
index 574da5bd69..6528ea2841 100644
--- a/lib-python/3/test/test_urlparse.py
+++ b/lib-python/3/test/test_urlparse.py
@@ -724,6 +724,24 @@ class UrlParseTestCase(unittest.TestCase):
with self.assertRaises(ValueError):
p.port
+ def test_attributes_bad_scheme(self):
+ """Check handling of invalid schemes."""
+ for bytes in (False, True):
+ for parse in (urllib.parse.urlsplit, urllib.parse.urlparse):
+ for scheme in (".", "+", "-", "0", "http&", "६http"):
+ with self.subTest(bytes=bytes, parse=parse, scheme=scheme):
+ url = scheme + "://www.example.net"
+ if bytes:
+ if url.isascii():
+ url = url.encode("ascii")
+ else:
+ continue
+ p = parse(url)
+ if bytes:
+ self.assertEqual(p.scheme, b"")
+ else:
+ self.assertEqual(p.scheme, "")
+
def test_attributes_without_netloc(self):
# This example is straight from RFC 3261. It looks like it
# should allow the username, hostname, and port to be filled
diff --git a/lib-python/3/urllib/parse.py b/lib-python/3/urllib/parse.py
index f5d3662313..1bb2f5b4a7 100644
--- a/lib-python/3/urllib/parse.py
+++ b/lib-python/3/urllib/parse.py
@@ -482,7 +482,7 @@ def urlsplit(url, scheme='', allow_fragments=True):
clear_cache()
netloc = query = fragment = ''
i = url.find(':')
- if i > 0:
+ if i > 0 and url[0].isascii() and url[0].isalpha():
for c in url[:i]:
if c not in scheme_chars:
break