[utils] strip #HttpOnly_ prefix from cookies files (#20219)

test/test_YoutubeDLCookieJar.py

@@ -29,6 +29,16 @@ def test_keep_session_cookies(self):
             tf.close()
             os.remove(tf.name)
 
+    def test_strip_httponly_prefix(self):
+        cookiejar = YoutubeDLCookieJar('./test/testdata/cookies/httponly_cookies.txt')
+        cookiejar.load(ignore_discard=True, ignore_expires=True)
+
+        def assert_cookie_has_value(key):
+            self.assertEqual(cookiejar._cookies['www.foobar.foobar']['/'][key].value, key + '_VALUE')
+
+        assert_cookie_has_value('HTTPONLY_COOKIE')
+        assert_cookie_has_value('JS_ACCESSIBLE_COOKIE')
+
 
 if __name__ == '__main__':
     unittest.main()

test/testdata/cookies/httponly_cookies.txt

@@ -0,0 +1,6 @@
+# Netscape HTTP Cookie File
+# http://curl.haxx.se/rfc/cookie_spec.html
+# This is a generated file!  Do not edit.
+
+#HttpOnly_www.foobar.foobar	FALSE	/	TRUE	2147483647	HTTPONLY_COOKIE	HTTPONLY_COOKIE_VALUE
+www.foobar.foobar	FALSE	/	TRUE	2147483647	JS_ACCESSIBLE_COOKIE	JS_ACCESSIBLE_COOKIE_VALUE

youtube_dl/utils.py

@@ -1141,6 +1141,8 @@ def https_open(self, req):
 
 
 class YoutubeDLCookieJar(compat_cookiejar.MozillaCookieJar):
+    _HTTPONLY_PREFIX = '#HttpOnly_'
+
     def save(self, filename=None, ignore_discard=False, ignore_expires=False):
         # Store session cookies with `expires` set to 0 instead of an empty
         # string
@@ -1150,7 +1152,21 @@ def save(self, filename=None, ignore_discard=False, ignore_expires=False):
         compat_cookiejar.MozillaCookieJar.save(self, filename, ignore_discard, ignore_expires)
 
     def load(self, filename=None, ignore_discard=False, ignore_expires=False):
-        compat_cookiejar.MozillaCookieJar.load(self, filename, ignore_discard, ignore_expires)
+        """Load cookies from a file."""
+        if filename is None:
+            if self.filename is not None:
+                filename = self.filename
+            else:
+                raise ValueError(compat_cookiejar.MISSING_FILENAME_TEXT)
+
+        cf = io.StringIO()
+        with open(filename) as f:
+            for line in f:
+                if line.startswith(self._HTTPONLY_PREFIX):
+                    line = line[len(self._HTTPONLY_PREFIX):]
+                cf.write(compat_str(line))
+        cf.seek(0)
+        self._really_load(cf, filename, ignore_discard, ignore_expires)
         # Session cookies are denoted by either `expires` field set to
         # an empty string or 0. MozillaCookieJar only recognizes the former
         # (see [1]). So we need force the latter to be recognized as session