Replace inline cloning with option-pop hook for <selectedcontent>

The initial implementation did inline cloning during parsing (mirroring
elements/characters into <selectedcontent> as tokens were processed) AND
deep-cloned from the DOM when option was popped.

The inline cloning was wrong per spec — cloning should operate on the
DOM, not on what the parser saw (the adoption agency algorithm can
restructure the tree) — and redundant, since pop() cleared
<selectedcontent> and deep-cloned from the DOM anyway.

This replaces all inline cloning machinery with a single overridable
hook method, cloneOptionContentToSelectedContent(), called when <option>
is popped. The tree builder tracks which <option> is active and where
<selectedcontent> is, but delegates the actual DOM cloning to subclasses.

Author: sideshowbarker

src/nu/validator/htmlparser/impl/TreeBuilder.java

@@ -35,7 +35,6 @@
 
 package nu.validator.htmlparser.impl;
 
-import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.HashMap;
 import java.util.Map;
@@ -444,10 +443,6 @@ public abstract class TreeBuilder<T> implements TokenHandler,
     // Tracks if we've already had an active option (first option was selected for cloning)
     private boolean hadActiveOption = false;
 
-    // Stack to track the current parent in selectedcontent for element cloning
-    // When we're inside an active option, we push cloned elements here
-    private ArrayList<T> selectedContentCloneStack = new ArrayList<T>();
-
     protected @Auto char[] charBuffer;
 
     protected int charBufferLen = 0;
@@ -2355,10 +2350,6 @@ public final void startTag(ElementName elementName,
                                     if (selectedContentPointer != null) {
                                         if (hasSelected) {
                                             // Option with selected attr becomes active
-                                            // Clear previous selectedcontent content if we had a different active option
-                                            if (hadActiveOption && !seenSelectedOption) {
-                                                clearSelectedContentChildren();
-                                            }
                                             seenSelectedOption = true;
                                             shouldBeActive = true;
                                         } else if (!seenSelectedOption && !hadActiveOption) {
@@ -2372,9 +2363,6 @@ public final void startTag(ElementName elementName,
                                     if (shouldBeActive) {
                                         activeOptionStackPos = currentPtr;
                                         hadActiveOption = true;
-                                        // Initialize the clone stack with selectedcontent as the root parent
-                                        selectedContentCloneStack.clear();
-                                        selectedContentCloneStack.add(selectedContentPointer);
                                     }
                                     attributes = null; // CPP
                                     break starttagloop;
@@ -5338,14 +5326,9 @@ private void pop() throws SAXException {
         // This handles adoption agency restructuring correctly
         if (currentPtr == activeOptionStackPos) {
             if (selectedContentPointer != null) {
-                clearSelectedContentChildren();
-                deepCloneChildren(node.node, selectedContentPointer);
+                cloneOptionContentToSelectedContent(node.node, selectedContentPointer);
             }
             activeOptionStackPos = -1;
-            selectedContentCloneStack.clear();
-        } else if (activeOptionStackPos >= 0 && currentPtr > activeOptionStackPos && selectedContentCloneStack.size() > 1) {
-            // Pop from clone stack when popping an element inside active option
-            selectedContentCloneStack.remove(selectedContentCloneStack.size() - 1);
         }
         // Clear selectedcontent tracking if we're popping the select element
         // (not when popping selectedcontent itself - the DOM node is still valid)
@@ -5355,7 +5338,6 @@ private void pop() throws SAXException {
             seenSelectedOption = false;
             activeOptionStackPos = -1;
             hadActiveOption = false;
-            selectedContentCloneStack.clear();
         }
         currentPtr--;
         elementPopped(node.ns, node.popName, node.node);
@@ -5371,14 +5353,9 @@ private void popForeign(int origPos, int eltPos) throws SAXException {
         // When active option closes, deep-clone its content to selectedcontent
         if (currentPtr == activeOptionStackPos) {
             if (selectedContentPointer != null) {
-                clearSelectedContentChildren();
-                deepCloneChildren(node.node, selectedContentPointer);
+                cloneOptionContentToSelectedContent(node.node, selectedContentPointer);
             }
             activeOptionStackPos = -1;
-            selectedContentCloneStack.clear();
-        } else if (activeOptionStackPos >= 0 && currentPtr > activeOptionStackPos && selectedContentCloneStack.size() > 1) {
-            // Pop from clone stack when popping an element inside active option
-            selectedContentCloneStack.remove(selectedContentCloneStack.size() - 1);
         }
         // Clear selectedcontent tracking if we're popping the select element
         if (node.getGroup() == SELECT) {
@@ -5387,7 +5364,6 @@ private void popForeign(int origPos, int eltPos) throws SAXException {
             seenSelectedOption = false;
             activeOptionStackPos = -1;
             hadActiveOption = false;
-            selectedContentCloneStack.clear();
         }
         currentPtr--;
         elementPopped(node.ns, node.popName, node.node);
@@ -5400,14 +5376,9 @@ private void silentPop() throws SAXException {
         // When active option closes, deep-clone its content to selectedcontent
         if (currentPtr == activeOptionStackPos) {
             if (selectedContentPointer != null) {
-                clearSelectedContentChildren();
-                deepCloneChildren(node.node, selectedContentPointer);
+                cloneOptionContentToSelectedContent(node.node, selectedContentPointer);
             }
             activeOptionStackPos = -1;
-            selectedContentCloneStack.clear();
-        } else if (activeOptionStackPos >= 0 && currentPtr > activeOptionStackPos && selectedContentCloneStack.size() > 1) {
-            // Pop from clone stack when popping an element inside active option
-            selectedContentCloneStack.remove(selectedContentCloneStack.size() - 1);
         }
         // Clear selectedcontent tracking if we're popping the select element
         if (node.getGroup() == SELECT) {
@@ -5416,7 +5387,6 @@ private void silentPop() throws SAXException {
             seenSelectedOption = false;
             activeOptionStackPos = -1;
             hadActiveOption = false;
-            selectedContentCloneStack.clear();
         }
         currentPtr--;
         node.release(this);
@@ -5428,14 +5398,9 @@ private void popOnEof() throws SAXException {
         // When active option closes, deep-clone its content to selectedcontent
         if (currentPtr == activeOptionStackPos) {
             if (selectedContentPointer != null) {
-                clearSelectedContentChildren();
-                deepCloneChildren(node.node, selectedContentPointer);
+                cloneOptionContentToSelectedContent(node.node, selectedContentPointer);
             }
             activeOptionStackPos = -1;
-            selectedContentCloneStack.clear();
-        } else if (activeOptionStackPos >= 0 && currentPtr > activeOptionStackPos && selectedContentCloneStack.size() > 1) {
-            // Pop from clone stack when popping an element inside active option
-            selectedContentCloneStack.remove(selectedContentCloneStack.size() - 1);
         }
         // Clear selectedcontent tracking if we're popping the select element
         if (node.getGroup() == SELECT) {
@@ -5444,7 +5409,6 @@ private void popOnEof() throws SAXException {
             seenSelectedOption = false;
             activeOptionStackPos = -1;
             hadActiveOption = false;
-            selectedContentCloneStack.clear();
         }
         currentPtr--;
         markMalformedIfScript(node.node);
@@ -5635,8 +5599,6 @@ private void appendToCurrentNodeAndPushFormattingElementMayFoster(
         // ]NOCPP]
         // This method can't be called for custom elements
         HtmlAttributes clone = attributes.cloneAttributes();
-        // Clone to selectedcontent if inside active option (must be before createElement due to C++ attribute ownership)
-        cloneElementToSelectedContent("http://www.w3.org/1999/xhtml", elementName.getName(), attributes);
         // Attributes must not be read after calling createElement, because
         // createElement may delete attributes in C++.
         T elt;
@@ -5686,8 +5648,6 @@ private void appendToCurrentNodeAndPushElement(ElementName elementName,
         } else {
             appendElement(elt, currentNode);
         }
-        // Clone to selectedcontent if inside active option
-        cloneElementToSelectedContent("http://www.w3.org/1999/xhtml", elementName.getName(), attributes);
         StackNode<T> node = createStackNode(elementName, elt
                 // [NOCPP[
                 , errorHandler == null ? null : new TaintableLocatorImpl(tokenizer)
@@ -5720,8 +5680,6 @@ private void appendToCurrentNodeAndPushElementMayFoster(ElementName elementName,
                     );
             appendElement(elt, currentNode);
         }
-        // Clone to selectedcontent if inside active option
-        cloneElementToSelectedContent("http://www.w3.org/1999/xhtml", popName, attributes);
         StackNode<T> node = createStackNode(elementName, elt, popName
                 // [NOCPP[
                 , errorHandler == null ? null : new TaintableLocatorImpl(tokenizer)
@@ -5745,8 +5703,6 @@ private void appendToCurrentNodeAndPushElementMayFosterMathML(
                 && annotationXmlEncodingPermitsHtml(attributes)) {
             markAsHtmlIntegrationPoint = true;
         }
-        // Clone to selectedcontent if inside active option (must be before createElement due to C++ attribute ownership)
-        cloneElementToSelectedContent("http://www.w3.org/1998/Math/MathML", popName, attributes);
         // Attributes must not be read after calling createElement(), since
         // createElement may delete the object in C++.
         T elt;
@@ -5812,8 +5768,6 @@ private void appendToCurrentNodeAndPushElementMayFosterSVG(
             popName = checkPopName(popName);
         }
         // ]NOCPP]
-        // Clone to selectedcontent if inside active option
-        cloneElementToSelectedContent("http://www.w3.org/2000/svg", popName, attributes);
         T elt;
         StackNode<T> current = stack[currentPtr];
         if (current.isFosterParenting()) {
@@ -5843,8 +5797,6 @@ private void appendToCurrentNodeAndPushElementMayFoster(ElementName elementName,
         checkAttributes(attributes, "http://www.w3.org/1999/xhtml");
         // ]NOCPP]
         // Can't be called for custom elements
-        // Clone to selectedcontent if inside active option
-        cloneElementToSelectedContent("http://www.w3.org/1999/xhtml", elementName.getName(), attributes);
         T elt;
         T formOwner = form == null || fragment || isTemplateContents() ? null : form;
         StackNode<T> current = stack[currentPtr];
@@ -6033,55 +5985,6 @@ private void appendVoidFormToCurrent(HtmlAttributes attributes) throws SAXExcept
         elementPopped("http://www.w3.org/1999/xhtml", "form", elt);
     }
 
-    /**
-     * Clones an element to the selectedcontent hierarchy when inside an active option.
-     * This is used for the customizable select feature.
-     *
-     * @param ns The namespace URI
-     * @param name The element name
-     * @param attributes The attributes (will be cloned)
-     */
-    private void cloneElementToSelectedContent(@NsUri String ns, @Local String name,
-            HtmlAttributes attributes) throws SAXException {
-        if (activeOptionStackPos < 0 || selectedContentCloneStack.isEmpty()) {
-            return;
-        }
-        // Clone the attributes
-        HtmlAttributes clonedAttrs = attributes.cloneAttributes();
-        // Get the current parent in the selectedcontent hierarchy
-        T selectedContentParent = selectedContentCloneStack.get(selectedContentCloneStack.size() - 1);
-        // Create a clone element
-        T clone = createElement(ns, name, clonedAttrs, selectedContentParent
-                // CPPONLY: , htmlCreator(null)
-                );
-        // Append the clone to the selectedcontent parent
-        appendElement(clone, selectedContentParent);
-        // Push the clone to the stack so nested content goes into it
-        selectedContentCloneStack.add(clone);
-    }
-
-    /**
-     * Clones a void element to the selectedcontent hierarchy when inside an active option.
-     * Void elements don't need to be pushed to the stack since they have no children.
-     */
-    private void cloneVoidElementToSelectedContent(@NsUri String ns, @Local String name,
-            HtmlAttributes attributes) throws SAXException {
-        if (activeOptionStackPos < 0 || selectedContentCloneStack.isEmpty()) {
-            return;
-        }
-        // Clone the attributes
-        HtmlAttributes clonedAttrs = attributes.cloneAttributes();
-        // Get the current parent in the selectedcontent hierarchy
-        T selectedContentParent = selectedContentCloneStack.get(selectedContentCloneStack.size() - 1);
-        // Create a clone element
-        T clone = createElement(ns, name, clonedAttrs, selectedContentParent
-                // CPPONLY: , htmlCreator(null)
-                );
-        // Append the clone to the selectedcontent parent
-        appendElement(clone, selectedContentParent);
-        // Void elements don't need to be pushed to the stack
-    }
-
     // [NOCPP[
 
     private final void accumulateCharactersForced(@Const @NoLength char[] buf,
@@ -6117,10 +6020,6 @@ private final void accumulateCharactersForced(@Const @NoLength char[] buf,
     protected void accumulateCharacters(@Const @NoLength char[] buf, int start,
             int length) throws SAXException {
         appendCharacters(stack[currentPtr].node, buf, start, length);
-        // Also clone to selectedcontent if in active option
-        if (activeOptionStackPos >= 0 && !selectedContentCloneStack.isEmpty()) {
-            appendCharacters(selectedContentCloneStack.get(selectedContentCloneStack.size() - 1), buf, start, length);
-        }
     }
 
     // ------------------------------- //
@@ -6149,12 +6048,14 @@ protected abstract T createHtmlElementSetAsRoot(HtmlAttributes attributes)
     protected abstract void detachFromParent(T element) throws SAXException;
 
     /**
-     * Deep clones the children of the source element to the destination element.
-     * Used for cloning option content to selectedcontent.
-     * Default implementation does nothing. Subclasses should override.
+     * Called when the active option is popped from the stack, to clone
+     * the option's children into selectedcontent. Subclasses that support
+     * DOM operations should override this to clear selectedcontent and
+     * deep-clone the option's children into it.
      */
-    protected void deepCloneChildren(T source, T destination) throws SAXException {
-        // Default implementation does nothing
+    protected void cloneOptionContentToSelectedContent(T option, T selectedContent)
+            throws SAXException {
+        // Default: no-op (streaming SAX mode ignores cloning)
     }
 
     protected abstract boolean hasChildren(T element) throws SAXException;
@@ -6200,16 +6101,6 @@ protected abstract void appendCommentToDocument(@NoLength char[] buf,
     protected abstract void addAttributesToElement(T element,
             HtmlAttributes attributes) throws SAXException;
 
-    /**
-     * Clears all children from the selectedcontent element.
-     * Used when an option with 'selected' attribute is seen after
-     * content has already been cloned from a previous option.
-     */
-    protected void clearSelectedContentChildren() throws SAXException {
-        // Default implementation does nothing. Subclasses that support
-        // selectedcontent cloning can override this.
-    }
-
     protected void markMalformedIfScript(T elt) throws SAXException {
 
     }
@@ -6415,10 +6306,6 @@ && charBufferContainsNonWhitespace()) {
                     // reconstructing gave us a new current node
                     appendCharacters(currentNode(), charBuffer, 0,
                             charBufferLen);
-                    // Also clone to selectedcontent if in active option
-                    if (activeOptionStackPos >= 0 && !selectedContentCloneStack.isEmpty()) {
-                        appendCharacters(selectedContentCloneStack.get(selectedContentCloneStack.size() - 1), charBuffer, 0, charBufferLen);
-                    }
                     charBufferLen = 0;
                     return;
                 }
@@ -6428,29 +6315,17 @@ && charBufferContainsNonWhitespace()) {
 
                 if (templatePos >= tablePos) {
                     appendCharacters(stack[templatePos].node, charBuffer, 0, charBufferLen);
-                    // Also clone to selectedcontent if in active option
-                    if (activeOptionStackPos >= 0 && !selectedContentCloneStack.isEmpty()) {
-                        appendCharacters(selectedContentCloneStack.get(selectedContentCloneStack.size() - 1), charBuffer, 0, charBufferLen);
-                    }
                     charBufferLen = 0;
                     return;
                 }
 
                 StackNode<T> tableElt = stack[tablePos];
                 insertFosterParentedCharacters(charBuffer, 0, charBufferLen,
                         tableElt.node, stack[tablePos - 1].node);
-                // Also clone to selectedcontent if in active option
-                if (activeOptionStackPos >= 0 && !selectedContentCloneStack.isEmpty()) {
-                    appendCharacters(selectedContentCloneStack.get(selectedContentCloneStack.size() - 1), charBuffer, 0, charBufferLen);
-                }
                 charBufferLen = 0;
                 return;
             }
             appendCharacters(currentNode(), charBuffer, 0, charBufferLen);
-            // Also clone to selectedcontent if in active option
-            if (activeOptionStackPos >= 0 && !selectedContentCloneStack.isEmpty()) {
-                appendCharacters(selectedContentCloneStack.get(selectedContentCloneStack.size() - 1), charBuffer, 0, charBufferLen);
-            }
             charBufferLen = 0;
         }
     }

src/nu/validator/htmlparser/sax/SAXTreeBuilder.java

@@ -200,14 +200,13 @@ private Node previousSibling(Node table) {
     }
 
     @Override
-    protected void clearSelectedContentChildren() throws SAXException {
-        if (selectedContentPointer != null) {
-            ((ParentNode) selectedContentPointer).clearChildren();
-        }
+    protected void cloneOptionContentToSelectedContent(Element option, Element selectedContent)
+            throws SAXException {
+        ((ParentNode) selectedContent).clearChildren();
+        deepCloneChildren(option, selectedContent);
     }
 
-    @Override
-    protected void deepCloneChildren(Element source, Element destination) throws SAXException {
+    private void deepCloneChildren(Element source, Element destination) throws SAXException {
         Node child = source.getFirstChild();
         while (child != null) {
             deepCloneNode(child, destination);