pythongh-99418: Make urllib.parse.urlparse enforce that a scheme must begin with an alphabetical ASCII character. (pythonGH-99421)

miss-islington · kenballus · mgorny · commit 5355793b8649 · 2024-03-19T18:38:49.000+01:00
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 439b9cf) Co-authored-by: Ben Kallus <49924171+kenballus@users.noreply.github.com>
diff --git a/Lib/test/test_urlparse.py b/Lib/test/test_urlparse.py
@@ -735,6 +735,24 @@ def test_attributes_bad_port(self):
                         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/urllib/parse.py b/Lib/urllib/parse.py
@@ -452,7 +452,7 @@ 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():
         if url[:i] == 'http': # optimize the common case
             url = url[i+1:]
             if url[:2] == '//':
diff --git a/Misc/NEWS.d/next/Library/2022-11-12-15-45-51.gh-issue-99418.FxfAXS.rst b/Misc/NEWS.d/next/Library/2022-11-12-15-45-51.gh-issue-99418.FxfAXS.rst
@@ -0,0 +1,2 @@
+Fix bug in :func:`urllib.parse.urlparse` that causes URL schemes that begin
+with a digit, a plus sign, or a minus sign to be parsed incorrectly.

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,2 @@`
	`1`	+Fix bug in :func:`urllib.parse.urlparse` that causes URL schemes that begin
	`2`	`+with a digit, a plus sign, or a minus sign to be parsed incorrectly.`