web feed polling: if we get Content-Type application/xml, use the lin…

…k's type

fixes eg https://simonwillison.net/atom/everything/

tests/test_web.py

@@ -1901,6 +1901,31 @@ def test_poll_feed_rss(self, mock_create_task, mock_get, _):
         self.assert_task(mock_create_task, 'poll-feed', '/queue/poll-feed',
                          domain='user.com', eta_seconds=expected_eta)
 
+    @patch('oauth_dropins.webutil.appengine_config.tasks_client.create_task')
+    def test_poll_feed_xml_content_type(self, mock_create_task, mock_get, _):
+        common.RUN_TASKS_INLINE = False
+        self.user.obj.mf2 = ACTOR_MF2_REL_FEED_URL
+        self.user.obj.put()
+
+        feed = """\
+<?xml version="1.0" encoding="UTF-8"?>
+<entry xmlns="http://www.w3.org/2005/Atom">
+<uri>https://user.com/post</uri>
+<content>I hereby ☕ post</content>
+</entry>
+"""
+        mock_get.return_value = requests_response(
+            feed, headers={'Content-Type': 'application/xml; charset=utf-8'})
+
+        got = self.post('/queue/poll-feed', data={'domain': 'user.com'})
+        self.assertEqual(200, got.status_code)
+        self.assertEqual(NOW, self.user.key.get().last_polled_feed)
+
+        mock_get.assert_has_calls((
+            self.req('https://foo/atom'),
+        ))
+        assert Object.get_by_id('https://user.com/post')
+
     def test_poll_feed_fails(self, mock_get, _):
         common.RUN_TASKS_INLINE = False
         self.user.obj.mf2 = {

web.py

@@ -59,6 +59,7 @@
 FEED_TYPES = {
     atom.CONTENT_TYPE.split(';')[0]: 'atom',
     rss.CONTENT_TYPE.split(';')[0]: 'rss',
+    'application/xml': 'xml',
 }
 MIN_FEED_POLL_PERIOD = timedelta(hours=2)
 MAX_FEED_POLL_PERIOD = timedelta(weeks=1)
@@ -628,8 +629,8 @@ def poll_feed_task():
 
     # discover feed URL
     for url, info in user.obj.mf2.get('rel-urls', {}).items():
-        if ('alternate' in info.get('rels', [])
-                and info.get('type', '').split(';')[0] in FEED_TYPES.keys()):
+        rel_type = FEED_TYPES.get(info.get('type', '').split(';')[0])
+        if 'alternate' in info.get('rels', []) and rel_type:
             break
     else:
         msg = f"User {user.key.id()} has no feed URL, can't fetch feed"
@@ -640,10 +641,10 @@ def poll_feed_task():
     resp = util.requests_get(url)
     content_type = resp.headers.get('Content-Type')
     type = FEED_TYPES.get(content_type.split(';')[0])
-    if type == 'atom':
+    if type == 'atom' or (type == 'xml' and rel_type == 'atom'):
         activities = atom.atom_to_activities(resp.text)
         obj_feed_prop = {'atom': resp.text}
-    elif type == 'rss':
+    elif type == 'rss' or (type == 'xml' and rel_type == 'rss'):
         activities = rss.to_activities(resp.text)
         obj_feed_prop = {'rss': resp.text}
     else: