Updated resetInsertionMode

Fixes #1491 Need to add support for template tags
jhy · Jul 11, 2021 · 2676f41 · 2676f41
1 parent 049204c
commit 2676f41
Show file tree

Hide file tree

Showing 2 changed files with 17 additions and 10 deletions.
diff --git a/CHANGES b/CHANGES
@@ -5,6 +5,9 @@ jsoup changelog
 
   * Bugfix: when making a HTTP POST, if the request write fails, make sure the connection is immediately cleaned up.
 
+  * Bugfix: updated the HtmlTreeParser resetInsertionMode to the current spec for supported elements
+    <https://github.com/jhy/jsoup/issues/1491>
+
   * Bugfix [Fuzz]: fixed a slow parse when a tag has thousands of null characters in it.
     <https://github.com/jhy/jsoup/issues/1580>
 

diff --git a/src/main/java/org/jsoup/parser/HtmlTreeBuilder.java b/src/main/java/org/jsoup/parser/HtmlTreeBuilder.java
@@ -438,17 +438,20 @@ private void replaceInQueue(ArrayList<Element> queue, Element out, Element in) {
     }
 
     void resetInsertionMode() {
+        // https://html.spec.whatwg.org/multipage/parsing.html#the-insertion-mode
         boolean last = false;
         for (int pos = stack.size() -1; pos >= 0; pos--) {
             Element node = stack.get(pos);
             if (pos == 0) {
                 last = true;
-                node = contextElement;
+                if (fragmentParsing)
+                    node = contextElement;
             }
             String name = node != null ? node.normalName() : "";
             if ("select".equals(name)) {
                 transition(HtmlTreeBuilderState.InSelect);
-                break; // frag
+                // todo - should loop up (with some limit) and check for table or template hits
+                break;
             } else if (("td".equals(name) || "th".equals(name) && !last)) {
                 transition(HtmlTreeBuilderState.InCell);
                 break;
@@ -463,25 +466,26 @@ void resetInsertionMode() {
                 break;
             } else if ("colgroup".equals(name)) {
                 transition(HtmlTreeBuilderState.InColumnGroup);
-                break; // frag
+                break;
             } else if ("table".equals(name)) {
                 transition(HtmlTreeBuilderState.InTable);
                 break;
-            } else if ("head".equals(name)) {
-                transition(HtmlTreeBuilderState.InBody);
-                break; // frag
+            // todo - template
+            } else if ("head".equals(name) && !last) {
+                transition(HtmlTreeBuilderState.InHead);
+                break;
             } else if ("body".equals(name)) {
                 transition(HtmlTreeBuilderState.InBody);
                 break;
             } else if ("frameset".equals(name)) {
                 transition(HtmlTreeBuilderState.InFrameset);
-                break; // frag
+                break;
             } else if ("html".equals(name)) {
-                transition(HtmlTreeBuilderState.BeforeHead);
-                break; // frag
+                transition(headElement == null ? HtmlTreeBuilderState.BeforeHead : HtmlTreeBuilderState.AfterHead);
+                break;
             } else if (last) {
                 transition(HtmlTreeBuilderState.InBody);
-                break; // frag
+                break;
             }
         }
     }