SDAAP-123 Escaped Markup being included in headlines

superdesk · Sep 30, 2024 · 1cd9a67 · 1cd9a67
1 parent 416f873
commit 1cd9a67
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 1 deletion.
diff --git a/server/aap/publish/formatters/reuters_newsml_1_2_formatter.py b/server/aap/publish/formatters/reuters_newsml_1_2_formatter.py
@@ -20,6 +20,7 @@
 from flask import current_app as app
 from apps.prepopulate.app_initialize import get_filepath
 from superdesk import etree as sd_etree
+from superdesk.text_utils import get_text
 from copy import deepcopy
 from superdesk.utc import utcnow
 from superdesk.editor_utils import remove_all_embeds
@@ -277,7 +278,8 @@ def _format_news_lines(self, formatted_article, main_news_component):
         """
         news_lines = SubElement(main_news_component, "NewsLines")
         if formatted_article.get('headline'):
-            SubElement(news_lines, 'HeadLine').text = formatted_article.get('headline')
+            SubElement(news_lines, 'HeadLine').text = (
+                get_text(formatted_article.get('headline'), 'html'))
         if formatted_article.get('byline'):
             SubElement(news_lines, 'ByLine').text = formatted_article.get('byline') or ''
         if formatted_article.get('dateline', {}).get('text', ''):

diff --git a/server/aap/publish/formatters/reuters_newsml_1_2_formatter_test.py b/server/aap/publish/formatters/reuters_newsml_1_2_formatter_test.py
@@ -213,6 +213,7 @@ def test_embeded(self):
         self.article['versioncreated'] = now
         item = self.article.copy()
         item.update({
+            'headline': 'headline is here',
             'body_html':
                 '<p>pre amble</p>'
                 '<!-- EMBED START Image {id: \"editor_0\"} -->'
@@ -226,6 +227,8 @@ def test_embeded(self):
             'format': 'HTML',
             "fields_meta": {
                 "body_html": {
+                },
+                "headline": {
                 }
             }
         })
@@ -237,3 +240,5 @@ def test_embeded(self):
                          etree.tostring(newsml.find('./NewsItem/NewsComponent/NewsComponent/ContentItem/DataContent/'
                                                     '{http://www.w3.org/1999/xhtml}html/'
                                                     '{http://www.w3.org/1999/xhtml}body')).decode('utf-8'))
+        self.assertEqual('headline is here',
+                         newsml.find('./NewsItem/NewsComponent/NewsComponent/NewsLines/HeadLine').text)