LibHTML: Add Comment and CharacterData nodes and improve HTML parsing

This patch adds the CharacterData subclass of Node, which is now the parent class of Text and a new Comment class. A Comment node is one of these in HTML:  Since these occur somewhat frequently on the web, we need to be able to parse them. This patch also adds a child rejection mechanism to the DOM tree. Nodes can now override is_child_allowed(Node) and return false if they don't want a particular Node to become a child of theirs. This is used to prevent Document from taking on unwanted children.
author: Andreas Kling <awesomekling@gmail.com> 2019-10-12 23:26:47 +0200
committer: Andreas Kling <awesomekling@gmail.com> 2019-10-12 23:34:05 +0200
commit: b083a233d8482ed7106d7a102866ee0dd42acc94 (patch)
tree: d32b52a7cc55fcd66dcaa94bb51f01e4bd04f398
parent: 6d150df58a0b64c4cb583fc374fd05e60d387dc6 (diff)
download: serenity-b083a233d8482ed7106d7a102866ee0dd42acc94.zip
15 files changed, 158 insertions, 25 deletions
diff --git a/Base/home/anon/www/welcome.html b/Base/home/anon/www/welcome.html
index 24602f109e..5f6739bd16 100644
--- a/Base/home/anon/www/welcome.html
+++ b/Base/home/anon/www/welcome.html
@@ -1,6 +1,8 @@
+<!DOCTYPE html>
 <html>
 <head>
 <title>Welcome!</title>
+<!-- this is a comment -->
 <style type="text/css">
 body {
     background-color: #fff;
diff --git a/Libraries/LibHTML/DOM/CharacterData.cpp b/Libraries/LibHTML/DOM/CharacterData.cpp
new file mode 100644
index 0000000000..e4b9d74e13
--- /dev/null
+++ b/Libraries/LibHTML/DOM/CharacterData.cpp
@@ -0,0 +1,11 @@
+#include <LibHTML/DOM/CharacterData.h>
+
+CharacterData::CharacterData(Document& document, NodeType type, const String& data)
+    : Node(document, type)
+    , m_data(data)
+{
+}
+
+CharacterData::~CharacterData()
+{
+}
diff --git a/Libraries/LibHTML/DOM/CharacterData.h b/Libraries/LibHTML/DOM/CharacterData.h
new file mode 100644
index 0000000000..34170ef6b9
--- /dev/null
+++ b/Libraries/LibHTML/DOM/CharacterData.h
@@ -0,0 +1,25 @@
+#pragma once
+
+#include <AK/String.h>
+#include <LibHTML/DOM/Node.h>
+
+class CharacterData : public Node {
+public:
+    virtual ~CharacterData() override;
+
+    const String& data() const { return m_data; }
+
+    virtual String text_content() const override { return m_data; }
+
+protected:
+    explicit CharacterData(Document&, NodeType, const String&);
+
+private:
+    String m_data;
+};
+
+template<>
+inline bool is<CharacterData>(const Node& node)
+{
+    return node.is_character_data();
+}
diff --git a/Libraries/LibHTML/DOM/Comment.cpp b/Libraries/LibHTML/DOM/Comment.cpp
new file mode 100644
index 0000000000..3d97339e49
--- /dev/null
+++ b/Libraries/LibHTML/DOM/Comment.cpp
@@ -0,0 +1,11 @@
+#include <LibHTML/DOM/Comment.h>
+#include <LibHTML/Layout/LayoutText.h>
+
+Comment::Comment(Document& document, const String& data)
+    : CharacterData(document, NodeType::COMMENT_NODE, data)
+{
+}
+
+Comment::~Comment()
+{
+}
diff --git a/Libraries/LibHTML/DOM/Comment.h b/Libraries/LibHTML/DOM/Comment.h
new file mode 100644
index 0000000000..39d0cd4cad
--- /dev/null
+++ b/Libraries/LibHTML/DOM/Comment.h
@@ -0,0 +1,18 @@
+#pragma once
+
+#include <AK/String.h>
+#include <LibHTML/DOM/CharacterData.h>
+
+class Comment final : public CharacterData {
+public:
+    explicit Comment(Document&, const String&);
+    virtual ~Comment() override;
+
+    virtual String tag_name() const override { return "#comment"; }
+};
+
+template<>
+inline bool is<Comment>(const Node& node)
+{
+    return node.is_comment();
+}
diff --git a/Libraries/LibHTML/DOM/Document.cpp b/Libraries/LibHTML/DOM/Document.cpp
index 8b4af6a0bd..4262bca329 100644
--- a/Libraries/LibHTML/DOM/Document.cpp
+++ b/Libraries/LibHTML/DOM/Document.cpp
@@ -29,6 +29,23 @@ StyleResolver& Document::style_resolver()
     return *m_style_resolver;
 }
 
+bool Document::is_child_allowed(const Node& node) const
+{
+    switch (node.type()) {
+    case NodeType::DOCUMENT_NODE:
+    case NodeType::TEXT_NODE:
+        return false;
+    case NodeType::COMMENT_NODE:
+        return true;
+    case NodeType::DOCUMENT_TYPE_NODE:
+        return !first_child_of_type<DocumentType>();
+    case NodeType::ELEMENT_NODE:
+        return !first_child_of_type<Element>();
+    default:
+        return false;
+    }
+}
+
 void Document::fixup()
 {
     if (!is<DocumentType>(first_child()))
diff --git a/Libraries/LibHTML/DOM/Document.h b/Libraries/LibHTML/DOM/Document.h
index 485ed87cec..8f3947147c 100644
--- a/Libraries/LibHTML/DOM/Document.h
+++ b/Libraries/LibHTML/DOM/Document.h
@@ -67,6 +67,8 @@ public:
     void invalidate_layout();
     Function<void()> on_invalidate_layout;
 
+    virtual bool is_child_allowed(const Node&) const override;
+
 private:
     virtual RefPtr<LayoutNode> create_layout_node(const StyleResolver&, const StyleProperties* parent_style) const override;
 
diff --git a/Libraries/LibHTML/DOM/DocumentType.h b/Libraries/LibHTML/DOM/DocumentType.h
index c13bb5cbee..78f2eae5bf 100644
--- a/Libraries/LibHTML/DOM/DocumentType.h
+++ b/Libraries/LibHTML/DOM/DocumentType.h
@@ -7,7 +7,7 @@ public:
     explicit DocumentType(Document&);
     virtual ~DocumentType() override;
 
-    virtual String tag_name() const override { return "!DOCTYPE"; }
+    virtual String tag_name() const override { return "#doctype"; }
 };
 
 template<>
diff --git a/Libraries/LibHTML/DOM/Node.h b/Libraries/LibHTML/DOM/Node.h
index 132bbcd63a..9efd4bf962 100644
--- a/Libraries/LibHTML/DOM/Node.h
+++ b/Libraries/LibHTML/DOM/Node.h
@@ -10,6 +10,7 @@ enum class NodeType : unsigned {
     INVALID = 0,
     ELEMENT_NODE = 1,
     TEXT_NODE = 3,
+    COMMENT_NODE = 8,
     DOCUMENT_NODE = 9,
     DOCUMENT_TYPE_NODE = 10,
 };
@@ -32,6 +33,8 @@ public:
     bool is_text() const { return type() == NodeType::TEXT_NODE; }
     bool is_document() const { return type() == NodeType::DOCUMENT_NODE; }
     bool is_document_type() const { return type() == NodeType::DOCUMENT_TYPE_NODE; }
+    bool is_comment() const { return type() == NodeType::COMMENT_NODE; }
+    bool is_character_data() const { return type() == NodeType::TEXT_NODE || type() == NodeType::COMMENT_NODE; }
     bool is_parent_node() const { return is_element() || is_document(); }
 
     virtual RefPtr<LayoutNode> create_layout_node(const StyleResolver&, const StyleProperties* parent_style) const;
@@ -66,6 +69,8 @@ public:
     const Element* previous_element_sibling() const;
     const Element* next_element_sibling() const;
 
+    virtual bool is_child_allowed(const Node&) const { return true; }
+
 protected:
     Node(Document&, NodeType);
 
diff --git a/Libraries/LibHTML/DOM/Text.cpp b/Libraries/LibHTML/DOM/Text.cpp
index a066a767f6..4dc3d34ff9 100644
--- a/Libraries/LibHTML/DOM/Text.cpp
+++ b/Libraries/LibHTML/DOM/Text.cpp
@@ -2,8 +2,7 @@
 #include <LibHTML/Layout/LayoutText.h>
 
 Text::Text(Document& document, const String& data)
-    : Node(document, NodeType::TEXT_NODE)
-    , m_data(data)
+    : CharacterData(document, NodeType::TEXT_NODE, data)
 {
 }
 
diff --git a/Libraries/LibHTML/DOM/Text.h b/Libraries/LibHTML/DOM/Text.h
index c507dbb3ba..30dd7d02c8 100644
--- a/Libraries/LibHTML/DOM/Text.h
+++ b/Libraries/LibHTML/DOM/Text.h
@@ -1,23 +1,17 @@
 #pragma once
 
 #include <AK/String.h>
-#include <LibHTML/DOM/Node.h>
+#include <LibHTML/DOM/CharacterData.h>
 
-class Text final : public Node {
+class Text final : public CharacterData {
 public:
     explicit Text(Document&, const String&);
     virtual ~Text() override;
 
-    const String& data() const { return m_data; }
-
     virtual String tag_name() const override { return "#text"; }
 
-    virtual String text_content() const override { return m_data; }
-
 private:
     virtual RefPtr<LayoutNode> create_layout_node(const StyleResolver&, const StyleProperties* parent_style) const override;
-
-    String m_data;
 };
 
 template<>
diff --git a/Libraries/LibHTML/Dump.cpp b/Libraries/LibHTML/Dump.cpp
index 6533b58774..2857d43e31 100644
--- a/Libraries/LibHTML/Dump.cpp
+++ b/Libraries/LibHTML/Dump.cpp
@@ -1,5 +1,6 @@
 #include <AK/Utf8View.h>
 #include <LibHTML/CSS/StyleSheet.h>
+#include <LibHTML/DOM/Comment.h>
 #include <LibHTML/DOM/Document.h>
 #include <LibHTML/DOM/DocumentType.h>
 #include <LibHTML/DOM/Element.h>
@@ -27,6 +28,8 @@ void dump_tree(const Node& node)
         dbgprintf("\"%s\"\n", static_cast<const Text&>(node).data().characters());
     } else if (is<DocumentType>(node)) {
         dbgprintf("<!DOCTYPE>\n");
+    } else if (is<Comment>(node)) {
+        dbgprintf("<!--%s-->\n", to<Comment>(node).data().characters());
     }
     ++indent;
     if (is<ParentNode>(node)) {
diff --git a/Libraries/LibHTML/Makefile.shared b/Libraries/LibHTML/Makefile.shared
index 08fc3d5500..a3af6d90d2 100644
--- a/Libraries/LibHTML/Makefile.shared
+++ b/Libraries/LibHTML/Makefile.shared
@@ -17,6 +17,8 @@ LIBHTML_OBJS = \
     DOM/HTMLBlinkElement.o \
     DOM/HTMLBRElement.o \
     DOM/Document.o \
+    DOM/CharacterData.o \
+    DOM/Comment.o \
     DOM/Text.o \
     DOM/DocumentType.o \
     DOM/ElementFactory.o \
diff --git a/Libraries/LibHTML/Parser/HTMLParser.cpp b/Libraries/LibHTML/Parser/HTMLParser.cpp
index a69871c8a0..7f0a8c9c26 100644
--- a/Libraries/LibHTML/Parser/HTMLParser.cpp
+++ b/Libraries/LibHTML/Parser/HTMLParser.cpp
@@ -1,6 +1,7 @@
 #include <AK/Function.h>
 #include <AK/NonnullRefPtrVector.h>
 #include <AK/StringBuilder.h>
+#include <LibHTML/DOM/Comment.h>
 #include <LibHTML/DOM/DocumentType.h>
 #include <LibHTML/DOM/Element.h>
 #include <LibHTML/DOM/ElementFactory.h>
@@ -44,6 +45,8 @@ NonnullRefPtr<Document> parse_html(const StringView& html, const URL& url)
         Free = 0,
         BeforeTagName,
         InTagName,
+        InDoctype,
+        InComment,
         InAttributeList,
         InAttributeName,
         BeforeAttributeValue,
@@ -101,19 +104,16 @@ NonnullRefPtr<Document> parse_html(const StringView& html, const URL& url)
             close_tag();
     };
 
-    auto handle_exclamation_tag = [&] {
-        auto name = String::copy(tag_name_buffer);
-        tag_name_buffer.clear();
-        ASSERT(name == "DOCTYPE");
-        if (node_stack.size() != 1)
-            node_stack[node_stack.size() - 2].append_child(adopt(*new DocumentType(document)), false);
-        close_tag();
+    auto commit_doctype = [&] {
+        node_stack.last().append_child(adopt(*new DocumentType(document)), false);
+    };
+
+    auto commit_comment = [&] {
+        node_stack.last().append_child(adopt(*new Comment(document, text_buffer.to_string())), false);
     };
 
     auto commit_tag = [&] {
-        if (is_exclamation_tag)
-            handle_exclamation_tag();
-        else if (is_slash_tag)
+        if (is_slash_tag)
             close_tag();
         else
             open_tag();
@@ -124,12 +124,16 @@ NonnullRefPtr<Document> parse_html(const StringView& html, const URL& url)
     };
 
     for (int i = 0; i < html.length(); ++i) {
+        auto peek = [&](int offset) -> char {
+            if (i + offset >= html.length())
+                return '\0';
+            return html[i + offset];
+        };
         char ch = html[i];
         switch (state) {
         case State::Free:
             if (ch == '<') {
                 is_slash_tag = false;
-                is_exclamation_tag = false;
                 move_to_state(State::BeforeTagName);
                 break;
             }
@@ -165,7 +169,22 @@ NonnullRefPtr<Document> parse_html(const StringView& html, const URL& url)
                 break;
             }
             if (ch == '!') {
-                is_exclamation_tag = true;
+                if (peek(1) == 'D'
+                    && peek(2) == 'O'
+                    && peek(3) == 'C'
+                    && peek(4) == 'T'
+                    && peek(5) == 'Y'
+                    && peek(6) == 'P'
+                    && peek(7) == 'E') {
+                    i += 7;
+                    move_to_state(State::InDoctype);
+                    break;
+                }
+                if (peek(1) == '-' && peek(2) == '-') {
+                    i += 2;
+                    move_to_state(State::InComment);
+                    break;
+                }
                 break;
             }
             if (ch == '>') {
@@ -188,6 +207,22 @@ NonnullRefPtr<Document> parse_html(const StringView& html, const URL& url)
             }
             tag_name_buffer.append(ch);
             break;
+        case State::InDoctype:
+            if (ch == '>') {
+                commit_doctype();
+                move_to_state(State::Free);
+                break;
+            }
+            break;
+        case State::InComment:
+            if (ch == '-' && peek(1) == '-' && peek(2) == '>') {
+                commit_comment();
+                i += 2;
+                move_to_state(State::Free);
+                break;
+            }
+            text_buffer.append(ch);
+            break;
         case State::InAttributeList:
             if (ch == '>') {
                 commit_tag();
diff --git a/Libraries/LibHTML/TreeNode.h b/Libraries/LibHTML/TreeNode.h
index 75f878fd80..2b498720f4 100644
--- a/Libraries/LibHTML/TreeNode.h
+++ b/Libraries/LibHTML/TreeNode.h
@@ -50,8 +50,10 @@ public:
     void append_child(NonnullRefPtr<T> node, bool call_inserted_into = true);
     void donate_all_children_to(T& node);
 
+    bool is_child_allowed(const T&) const { return true; }
+
 protected:
-    TreeNode() { }
+    TreeNode() {}
 
 private:
     int m_ref_count { 1 };
@@ -66,6 +68,10 @@ template<typename T>
 inline void TreeNode<T>::append_child(NonnullRefPtr<T> node, bool call_inserted_into)
 {
     ASSERT(!node->m_parent);
+
+    if (!static_cast<T*>(this)->is_child_allowed(*node))
+        return;
+
     if (m_last_child)
         m_last_child->m_next_sibling = node.ptr();
     node->m_previous_sibling = m_last_child;
@@ -82,6 +88,10 @@ template<typename T>
 inline void TreeNode<T>::prepend_child(NonnullRefPtr<T> node, bool call_inserted_into)
 {
     ASSERT(!node->m_parent);
+
+    if (!static_cast<T*>(this)->is_child_allowed(*node))
+        return;
+
     if (m_first_child)
         m_first_child->m_previous_sibling = node.ptr();
     node->m_next_sibling = m_first_child;
@@ -112,7 +122,6 @@ inline void TreeNode<T>::donate_all_children_to(T& node)
     m_last_child = nullptr;
 }
 
-
 template<typename T>
 inline bool TreeNode<T>::is_ancestor_of(const TreeNode<T>& other) const
 {
author	Andreas Kling <awesomekling@gmail.com>	2019-10-12 23:26:47 +0200
committer	Andreas Kling <awesomekling@gmail.com>	2019-10-12 23:34:05 +0200
commit	b083a233d8482ed7106d7a102866ee0dd42acc94 (patch)
tree	d32b52a7cc55fcd66dcaa94bb51f01e4bd04f398
parent	6d150df58a0b64c4cb583fc374fd05e60d387dc6 (diff)
download	serenity-b083a233d8482ed7106d7a102866ee0dd42acc94.zip