atom.activities_to_atom: bug fix: avoid encoded </> characters in…

… title

fixes #629

README.md

@@ -378,6 +378,7 @@ _Non-breaking changes:_
     * Bug fix, emove incorrect `type="application/atom+xml"` from `rel="self"` `link` in `entry`.
     * Render `objectType: comment` attachments.
     * Remove invalid `<a>` element for tags.
+    * Bug fix: avoid encoded `<` and `>` characters in `title` ([#629](https://github.com/snarfed/granary/issues/629)).
   * Bug fixes in `activity_to_atom`/`activities_to_atom` for dict-valued `url` fields.
   * Render images in article/note attachments.
   * Render `objectType: service` attachments, eg Bluesky custom feeds.

granary/atom.py

@@ -377,6 +377,9 @@ def _prepare_activity(a, reader=True):
     a (dict): ActivityStreams 1 activity
     reader (bool): whether the output will be rendered in a feed reader.
       Currently just includes location if True, not otherwise.
+
+  Returns:
+    ``None``
   """
   act_type = as1.object_type(a)
   obj = as1.get_object(a) or a
@@ -390,16 +393,14 @@ def _prepare_activity(a, reader=True):
     # https://forum.newsblur.com/t/android-cant-read-line-pre-formatted-lines/6116
     white_space_pre=False))
 
-  # Make sure every activity has the title field, since Atom <entry> requires
-  # the title element.
-  if not a.get('title'):
-    a['title'] = util.ellipsize(_encode_ampersands(
-      a.get('displayName') or a.get('content') or obj.get('title') or
-      obj.get('displayName') or obj.get('content') or 'Untitled'))
-
-  # strip HTML tags. the Atom spec says title is plain text:
+  # Make sure every activity has displayName, since Atom <entry> requires the
+  # title element. and strip HTML tags, the Atom spec says title is plain text:
   # http://atomenabled.org/developers/syndication/#requiredEntryElements
-  a['title'] = xml.sax.saxutils.escape(util.parse_html(a['title']).get_text(''))
+  display_name = (a.get('displayName') or a.get('content') or obj.get('title')
+                  or obj.get('displayName') or obj.get('content') or 'Untitled')
+  a['displayName'] = util.ellipsize(xml.sax.saxutils.escape(
+    util.parse_html(display_name).get_text('')))
+
 
   children = []
   image_urls_seen = set()

granary/templates/_entry.atom

@@ -8,7 +8,7 @@
 
   {% set url = as1.get_url(activity) or as1.get_url(obj) %}
   <id>{{ activity.id or obj.id or url }}</id>
-  <title>{{ activity.title|safe }}</title>
+  <title>{{ activity.displayName|safe }}</title>
 
   {% if obj.summary %}
     <summary>{{ obj.summary }}</summary>

granary/tests/test_atom.py

@@ -361,10 +361,17 @@ def test_title(self):
 
   def test_strip_html_tags_from_titles(self):
     activity = copy.deepcopy(test_facebook.ACTIVITY)
-    activity['displayName'] = '<p>foo &amp; <a href="http://bar">bar</a></p>'
+    activity['content'] = '<p>foo &amp; <a href="http://bar">bar</a></p>'
     self.assert_multiline_in(
       '<title>foo &amp; bar</title>\n',
-      atom.activities_to_atom([activity], test_facebook.ACTOR))
+      atom.activities_to_atom([activity], {}))
+
+    # ellipsize in the middle of an HTML tag. (ellipsize defaults to 14 words)
+    del activity['displayName']
+    activity['content'] = '<p>I’ve been looking over Mike Hoerger’s <a href="https://www.pmc19.com/data/index.php">Pandemic Mitigation Collaborative - Data Tracker</a> which estimates and projects...</p>'
+    self.assert_multiline_in(
+      '<title>I’ve been looking over Mike Hoerger’s Pandemic Mitigation Collaborative - Data Tracker which estimates...</title>\n',
+      atom.activities_to_atom([activity], {}))
 
   def test_render_content_as_html(self):
     self.assert_multiline_in(