LibWeb: Make HTMLDocumentParser take an existing document

We shouldn't really be creating the document objects inside the parser,
since that makes it hard to hook up e.g JavaScript bindings early on.

Libraries/LibWeb/HTML/Parser/HTMLDocumentParser.cpp

@@ -112,21 +112,15 @@ static Vector<FlyString> s_quirks_public_ids = {
 
 RefPtr<DOM::Document> parse_html_document(const StringView& data, const URL& url, const String& encoding)
 {
-    HTMLDocumentParser parser(data, TextCodec::get_standardized_encoding(encoding));
+    auto document = DOM::Document::create(url);
+    HTMLDocumentParser parser(document, data, TextCodec::get_standardized_encoding(encoding));
     parser.run(url);
-    return parser.document();
+    return document;
 }
 
-HTMLDocumentParser::HTMLDocumentParser(const StringView& input, const String& encoding)
+HTMLDocumentParser::HTMLDocumentParser(DOM::Document& document, const StringView& input, const String& encoding)
     : m_tokenizer(input, encoding)
-{
-    m_document = DOM::Document::create();
-    m_document->set_encoding(encoding);
-}
-
-HTMLDocumentParser::HTMLDocumentParser(const StringView& input, const String& encoding, DOM::Document& existing_document)
-    : m_tokenizer(input, encoding)
-    , m_document(existing_document)
+    , m_document(document)
 {
     m_document->set_encoding(encoding);
 }
@@ -2975,7 +2969,8 @@ DOM::Document& HTMLDocumentParser::document()
 
 NonnullRefPtrVector<DOM::Node> HTMLDocumentParser::parse_html_fragment(DOM::Element& context_element, const StringView& markup)
 {
-    HTMLDocumentParser parser(markup, "utf-8");
+    auto temp_document = DOM::Document::create();
+    HTMLDocumentParser parser(*temp_document, markup, "utf-8");
     parser.m_context_element = context_element;
     parser.m_parsing_fragment = true;
     parser.document().set_quirks_mode(context_element.document().mode());
@@ -3022,5 +3017,4 @@ NonnullRefPtrVector<DOM::Node> HTMLDocumentParser::parse_html_fragment(DOM::Elem
     }
     return children;
 }
-
 }

Libraries/LibWeb/HTML/Parser/HTMLDocumentParser.h

@@ -63,8 +63,7 @@ RefPtr<DOM::Document> parse_html_document(const StringView&, const URL&, const S
 
 class HTMLDocumentParser {
 public:
-    HTMLDocumentParser(const StringView& input, const String& encoding);
-    HTMLDocumentParser(const StringView& input, const String& encoding, DOM::Document& existing_document);
+    HTMLDocumentParser(DOM::Document&, const StringView& input, const String& encoding);
     ~HTMLDocumentParser();
 
     void run(const URL&);
@@ -180,7 +179,7 @@ private:
     bool m_stop_parsing { false };
     size_t m_script_nesting_level { 0 };
 
-    RefPtr<DOM::Document> m_document;
+    NonnullRefPtr<DOM::Document> m_document;
     RefPtr<HTMLHeadElement> m_head_element;
     RefPtr<HTMLFormElement> m_form_element;
     RefPtr<DOM::Element> m_context_element;

Libraries/LibWeb/Loader/FrameLoader.cpp

@@ -214,7 +214,8 @@ bool FrameLoader::load(const URL& url, Type type)
 
 void FrameLoader::load_html(const StringView& html, const URL& url)
 {
-    HTML::HTMLDocumentParser parser(html, "utf-8");
+    auto document = DOM::Document::create(url);
+    HTML::HTMLDocumentParser parser(document, html, "utf-8");
     parser.run(url);
     frame().set_document(&parser.document());
 }

Userland/test-web.cpp

@@ -240,7 +240,7 @@ void TestRunner::run()
         Web::ResourceLoader::the().load_sync(
             page_to_load,
             [&](auto data, auto&) {
-                Web::HTML::HTMLDocumentParser parser(data, "utf-8", *m_page_view->document());
+                Web::HTML::HTMLDocumentParser parser(*m_page_view->document(), data, "utf-8");
                 parser.run(page_to_load);
             },
             [page_to_load](auto error) {
@@ -350,7 +350,8 @@ JSFileResult TestRunner::run_file_test(const String& test_path)
         page_to_load,
         [&](auto data, auto&) {
             // Create a new parser and immediately get its document to replace the old interpreter.
-            Web::HTML::HTMLDocumentParser parser(data, "utf-8");
+            auto document = Web::DOM::Document::create();
+            Web::HTML::HTMLDocumentParser parser(document, data, "utf-8");
             auto& new_interpreter = parser.document().interpreter();
 
             // Setup the test environment and call "__BeforeInitialPageLoad__"