[FIX] link_tracker: allow mock to parse HTML to find shortened links

Some incoming tests crash if parser is not the HTML parser when trying to find
a node given its ID. We also remove an unnecessary encapsulation, allowing
to avoid issues like embedding a body or head tag within a div.

Task ID-2508643
Prepares Task ID-27033 (support QWeb in templates)
Prepares Task ID-2377974 (clean trace and status management in mass mailing)
COM PR odoo/odoo#69461
ENT PR odoo/enterprise#17780

X-original-commit: a442057f161610e7869d0863acd067cbadd18195

addons/link_tracker/tests/common.py

@@ -22,7 +22,8 @@ class MockLinkTracker(common.BaseCase):
         self.addCleanup(link_tracker_title_patch.stop)
 
     def _get_href_from_anchor_id(self, body, anchor_id):
-        html = etree.fromstring(body)
+        """ Parse en html body to find the href of an element given its ID. """
+        html = etree.fromstring(body, parser=etree.HTMLParser())
         return html.xpath("//*[@id='%s']" % anchor_id)[0].attrib.get('href')
 
     def _get_tracker_from_short_url(self, short_url):
@@ -41,7 +42,7 @@ class MockLinkTracker(common.BaseCase):
         )
         """
         (anchor_id, url, is_shortened) = link_info
-        anchor_href = self._get_href_from_anchor_id("<div>%s</div>" % body, anchor_id)
+        anchor_href = self._get_href_from_anchor_id(body, anchor_id)
         if is_shortened:
             self.assertTrue('/r/' in anchor_href, '%s should be shortened: %s' % (anchor_id, anchor_href))
             link_tracker = self._get_tracker_from_short_url(anchor_href)