Handle </p> and </br> in foreign contexts

stevecheckoway · stevecheckoway · commit b317bb8f6dda · 2021-06-29T16:26:43.000-04:00
The HTML spec changed to handle `&lt;/p&gt;` and `&lt;/br&gt;` in foreign context
the same way `&lt;p&gt;` and `&lt;br&gt;` are.

The fragment case is handled identically to the nonfragment case now.
diff --git a/gumbo-parser/src/parser.c b/gumbo-parser/src/parser.c
@@ -4418,6 +4418,7 @@ static void handle_in_foreign_content(GumboParser* parser, GumboToken* token) {
         || token_has_attribute(token, "size")
       )
     )
+    || tag_in(token, kEndTag, &(const TagSet) { TAG(BR), TAG(P) })
   ) {
     /* Parse error */
     parser_add_parse_error(parser, token);
@@ -4427,20 +4428,17 @@ static void handle_in_foreign_content(GumboParser* parser, GumboToken* token) {
      * fragment parsing algorithm, then act as described in the "any other
      * start tag" entry below.
      */
-    if (!is_fragment_parser(parser)) {
-      do {
-        pop_current_node(parser);
-      } while (
-        !(
-          is_mathml_integration_point(get_current_node(parser))
-          || is_html_integration_point(get_current_node(parser))
-          || get_current_node(parser)->v.element.tag_namespace == GUMBO_NAMESPACE_HTML
-        )
-      );
-      parser->_parser_state->_reprocess_current_token = true;
-      return;
+    while (
+      !(
+        is_mathml_integration_point(get_current_node(parser))
+        || is_html_integration_point(get_current_node(parser))
+        || get_current_node(parser)->v.element.tag_namespace == GUMBO_NAMESPACE_HTML
+      )
+    ) {
+      pop_current_node(parser);
     }
-    // This is a start tag so the next if's then branch will be taken.
+    handle_in_body(parser, token);
+    return;
   }
 
   if (token->type == GUMBO_TOKEN_START_TAG) {
diff --git a/gumbo-parser/test/parser.cc b/gumbo-parser/test/parser.cc
@@ -2220,4 +2220,24 @@ TEST_F(GumboParserTest, FragmentWithoutForm) {
   EXPECT_EQ(0, GetChildCount(span));
 }
 
+TEST_F(GumboParserTest, ForeignFragment) {
+  ParseFragment("</p><foo>", "svg", GUMBO_NAMESPACE_SVG);
+  EXPECT_EQ(1, GetChildCount(root_));
+  GumboNode* html = GetChild(root_, 0);
+  ASSERT_EQ(GUMBO_NODE_ELEMENT, html->type);
+  EXPECT_EQ(GUMBO_TAG_HTML, html->v.element.tag);
+  EXPECT_EQ(2, GetChildCount(html));
+
+  ASSERT_EQ(2, GetChildCount(html));
+  GumboNode* p = GetChild(html, 0);
+  ASSERT_EQ(GUMBO_NODE_ELEMENT, p->type);
+  ASSERT_EQ(GUMBO_TAG_P, p->v.element.tag);
+  ASSERT_EQ(GUMBO_NAMESPACE_HTML, p->v.element.tag_namespace);
+
+  GumboNode* foo = GetChild(html, 1);
+  ASSERT_EQ(GUMBO_NODE_ELEMENT, foo->type);
+  ASSERT_EQ(std::string("foo"), foo->v.element.name);
+  ASSERT_EQ(GUMBO_NAMESPACE_SVG, foo->v.element.tag_namespace);
+}
+
 }  // namespace
