From 3a629f0fac71d84e486923c280f8a1d6765131a1 Mon Sep 17 00:00:00 2001
From: Julian Soreavis <263610811+soreavis@users.noreply.github.com>
Date: Sun, 19 Jul 2026 20:58:52 +0200
Subject: [PATCH] gh-44672: Stop pre-resolving the FTP host in urllib.request

FTPHandler.ftp_open() resolved the host with socket.gethostbyname()
before connecting.  gethostbyname() is IPv4-only, so an ftp:// URL for
an IPv6-only host failed there even when the server was reachable.

As diagnosed on the tracker in 2009, ftp_open() has no need to resolve
the host itself: it is passed to connect_ftp(), then to ftpwrapper, and
on to ftplib.FTP.connect().  That uses socket.create_connection(), which
resolves via getaddrinfo() and is IPv6-aware.

Remove the pre-resolution and let the host flow through unresolved.  An
unresolvable host now raises socket.gaierror inside ftplib; gaierror is
an OSError and therefore in ftplib.all_errors, so ftp_open()'s existing
handler still turns it into URLError and callers see the same error type
as before.
---
 Lib/test/test_urllib2.py                      | 41 ++++++++++++++++++-
 Lib/urllib/request.py                         |  4 --
 ...6-07-19-20-57-04.gh-issue-44672.nFrMYT.rst |  4 ++
 3 files changed, 44 insertions(+), 5 deletions(-)
 create mode 100644 Misc/NEWS.d/next/Library/2026-07-19-20-57-04.gh-issue-44672.nFrMYT.rst

diff --git a/Lib/test/test_urllib2.py b/Lib/test/test_urllib2.py
index d2fd111f6d9de02..2bec48a33f53bd8 100644
--- a/Lib/test/test_urllib2.py
+++ b/Lib/test/test_urllib2.py
@@ -787,7 +787,7 @@ def connect_ftp(self, user, passwd, host, port, dirs,
             # ftp authentication not yet implemented by FTPHandler
             self.assertEqual(h.user, user)
             self.assertEqual(h.passwd, passwd)
-            self.assertEqual(h.host, socket.gethostbyname(host))
+            self.assertEqual(h.host, host)
             self.assertEqual(h.port, port)
             self.assertEqual(h.dirs, dirs)
             self.assertEqual(h.ftpwrapper.filename, filename)
@@ -797,6 +797,45 @@ def connect_ftp(self, user, passwd, host, port, dirs,
             self.assertEqual(int(headers["Content-length"]), len(data))
             r.close()
 
+    def test_ftp_no_hostname_resolution(self):
+        # gh-44672: the host must reach connect_ftp unresolved so that
+        # ftplib resolves it via getaddrinfo(), which supports IPv6.
+        class MockFTPWrapper:
+            def retrfile(self, filename, filetype):
+                return io.StringIO(""), 0
+
+            def close(self):
+                pass
+
+        class NullFTPHandler(urllib.request.FTPHandler):
+            def connect_ftp(self, user, passwd, host, port, dirs, timeout):
+                self.host = host
+                return MockFTPWrapper()
+
+        h = NullFTPHandler()
+        req = Request("ftp://localhost/foo/bar.html")
+        req.timeout = None
+        with mock.patch.object(socket, "gethostbyname",
+                               side_effect=AssertionError(
+                                   "ftp_open must not pre-resolve the host")):
+            r = h.ftp_open(req)
+        r.close()
+        self.assertEqual(h.host, "localhost")
+
+    def test_ftp_gaierror(self):
+        # gh-44672: without pre-resolution the lookup failure now surfaces
+        # from ftplib as socket.gaierror; it must still become a URLError.
+        class GaiErrorFTPHandler(urllib.request.FTPHandler):
+            def connect_ftp(self, user, passwd, host, port, dirs, timeout):
+                raise socket.gaierror(-2, "Name or service not known")
+
+        h = GaiErrorFTPHandler()
+        req = Request("ftp://nonexistent.invalid/")
+        req.timeout = None
+        with self.assertRaises(urllib.error.URLError) as cm:
+            h.ftp_open(req)
+        self.assertIsInstance(cm.exception.__cause__, socket.gaierror)
+
     @support.requires_resource("network")
     def test_ftp_error(self):
         class ErrorFTPHandler(urllib.request.FTPHandler):
diff --git a/Lib/urllib/request.py b/Lib/urllib/request.py
index 660301fef612588..05696498987a085 100644
--- a/Lib/urllib/request.py
+++ b/Lib/urllib/request.py
@@ -1526,10 +1526,6 @@ def ftp_open(self, req):
         user = user or ''
         passwd = passwd or ''
 
-        try:
-            host = socket.gethostbyname(host)
-        except OSError as msg:
-            raise URLError(msg)
         path, attrs = _splitattr(req.selector)
         dirs = path.split('/')
         dirs = list(map(unquote, dirs))
diff --git a/Misc/NEWS.d/next/Library/2026-07-19-20-57-04.gh-issue-44672.nFrMYT.rst b/Misc/NEWS.d/next/Library/2026-07-19-20-57-04.gh-issue-44672.nFrMYT.rst
new file mode 100644
index 000000000000000..a74488e0b02ccc3
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2026-07-19-20-57-04.gh-issue-44672.nFrMYT.rst
@@ -0,0 +1,4 @@
+:func:`urllib.request.urlopen` can now fetch ``ftp://`` URLs whose host is
+IPv6-only.  :class:`~urllib.request.FTPHandler` no longer pre-resolves the host
+with the IPv4-only :func:`socket.gethostbyname`; the unresolved host is passed to
+:mod:`ftplib`, which resolves it with :func:`socket.getaddrinfo`.