Added additional entry point

• Added an option to use Document::loadHTML or Document::load to parse a document. • Made the DOM elements use dW\HTML5 namespace instead of dW\HTML5\DOM. • Fixed where TreeBuilder wasn't being properly destructed when the parser is finished.
6 years ago · fd6003fb4e
13 changed files with 71 additions and 21 deletions
--- a/README.md
+++ b/README.md
@ -6,4 +6,14 @@ Tools for parsing and printing HTML5 documents and fragments.
 <?php
 $dom = dW\HTML5\Parser::parse('<!DOCTYPE html><html lang="en" charset="utf-8"><head><title>Ook!</title></head><body><h1>Ook!</h1><p>Ook-ook? Oooook. Ook ook oook ook oooooook ook ooook ook.</p><p>Eek!</p></body></html>');
 ?>
-```
+```
+
+or:
+
+```php
+<?php
+$dom = new dW\HTML\Document;
+$dom->loadHTML('<!DOCTYPE html><html lang="en" charset="utf-8"><head><title>Ook!</title></head><body><h1>Ook!</h1><p>Ook-ook? Oooook. Ook ook oook ook oooooook ook ooook ook.</p><p>Eek!</p></body></html>');
+?>
+```
+
--- a/composer.json
+++ b/composer.json
@ -18,7 +18,10 @@
    ],
    "autoload": {
        "psr-4": {
-            "dW\\HTML5\\": "lib/"
+            "dW\\HTML5\\": [
+                "lib/",
+                "lib/DOM"
+            ]
        },
        "classmap": ["lib/Token.php"]
    },
--- a/lib/DOM/Comment.php
+++ b/lib/DOM/Comment.php
@ -1,6 +1,6 @@
 <?php
 declare(strict_types=1);
-namespace dW\HTML5\DOM;
+namespace dW\HTML5;

 class Comment extends \DOMComment {
    use Node;
--- a/lib/DOM/Document.php
+++ b/lib/DOM/Document.php
@ -1,6 +1,6 @@
 <?php
 declare(strict_types=1);
-namespace dW\HTML5\DOM;
+namespace dW\HTML5;

 class Document extends \DOMDocument {
    use Printer;
@ -8,9 +8,11 @@ class Document extends \DOMDocument {
    public function __construct() {
        parent::__construct();

-        $this->registerNodeClass('DOMComment', '\dW\HTML5\DOM\Comment');
-        $this->registerNodeClass('DOMElement', '\dW\HTML5\DOM\Element');
-        $this->registerNodeClass('DOMText', '\dW\HTML5\DOM\Text');
+        $this->registerNodeClass('DOMComment', '\dW\HTML5\Comment');
+        $this->registerNodeClass('DOMDocumentFragment', '\dW\HTML5\DocumentFragment');
+        $this->registerNodeClass('DOMElement', '\dW\HTML5\Element');
+        $this->registerNodeClass('DOMProcessingInstruction', '\dW\HTML5\ProcessingInstruction');
+        $this->registerNodeClass('DOMText', '\dW\HTML5\Text');
    }

    public function fixIdAttributes() {
@ -44,7 +46,17 @@ class Document extends \DOMDocument {
        $this->normalize();
    }

-    public function load($source, $options = null) {}
-    public function loadHTML($source, $options = null) {}
+    public function load($source, $options = null): bool {
+        Parser::parse((string)$source, $this, true);
+
+        return true;
+    }
+
+    public function loadHTML($source, $options = null): bool {
+        Parser::parse((string)$source, $this);
+
+        return true;
+    }
+
    public function loadXML($source, $options = null) {}
 }
--- a/lib/DOM/DocumentFragment.php
+++ b/lib/DOM/DocumentFragment.php
@ -0,0 +1,7 @@
+<?php
+declare(strict_types=1);
+namespace dW\HTML5;
+
+class DocumentFragment extends \DOMDocumentFragment {
+    use Node;
+}
--- a/lib/DOM/Element.php
+++ b/lib/DOM/Element.php
@ -1,6 +1,6 @@
 <?php
 declare(strict_types=1);
-namespace dW\HTML5\DOM;
+namespace dW\HTML5;

 class Element extends \DOMElement {
    use Node;
--- a/lib/DOM/Node.php
+++ b/lib/DOM/Node.php
@ -1,6 +1,6 @@
 <?php
 declare(strict_types=1);
-namespace dW\HTML5\DOM;
+namespace dW\HTML5;

 trait Node {
    public function getAncestor($needle): Element {
@ -37,7 +37,7 @@ trait Node {
                return $context;
            }
        } else {
-            throw new \dW\HTML5\Exception(\dW\HTML5\Exception::DOM_DOMNODE_STRING_OR_CLOSURE_EXPECTED, gettype($needle));
+            throw new Exception(Exception::DOM_DOMNODE_STRING_OR_CLOSURE_EXPECTED, gettype($needle));
        }

        return null;
--- a/lib/DOM/Printer.php
+++ b/lib/DOM/Printer.php
@ -1,6 +1,6 @@
 <?php
 declare(strict_types=1);
-namespace dW\HTML5\DOM;
+namespace dW\HTML5;

 trait Printer {
    protected $selfClosingElements = ['area', 'base', 'basefont', 'bgsound', 'br', 'col', 'embed', 'frame', 'hr', 'img', 'input', 'link', 'meta', 'param', 'source', 'track', 'wbr'];
@ -11,7 +11,7 @@ trait Printer {
        }

        if (!$node instanceof \DOMElement && !$node instanceof \DOMDocument && !$node instanceof \DOMDocumentFragment) {
-            throw new \dW\HTML5\Exception(\dW\HTML5\Exception::DOM_ELEMENT_DOCUMENT_DOCUMENTFRAG_EXPECTED, gettype($node));
+            throw new Exception(Exception::DOM_ELEMENT_DOCUMENT_DOCUMENTFRAG_EXPECTED, gettype($node));
        }

        # 1. Let s be a string, and initialize it to the empty string.
--- a/lib/DOM/ProcessingInstruction.php
+++ b/lib/DOM/ProcessingInstruction.php
@ -0,0 +1,7 @@
+<?php
+declare(strict_types=1);
+namespace dW\HTML5;
+
+class ProcessingInstruction extends \DOMProcessingInstruction {
+    use Node;
+}
--- a/lib/DOM/Text.php
+++ b/lib/DOM/Text.php
@ -1,6 +1,6 @@
 <?php
 declare(strict_types=1);
-namespace dW\HTML5\DOM;
+namespace dW\HTML5;

 class Text extends \DOMText {
    use Node;
--- a/lib/Exception.php
+++ b/lib/Exception.php
@ -21,6 +21,8 @@ class Exception extends \Exception {
    const TREEBUILDER_FORMELEMENT_EXPECTED = 10501;
    const TREEBUILDER_DOCUMENTFRAG_ELEMENT_DOCUMENT_DOCUMENTFRAG_EXPECTED = 10502;

+    const PARSER_NONEMPTY_DOCUMENT = 10601;
+
    protected static $messages = [10000 => 'Invalid error code',
                                  10001 => 'Unknown error; escaping',
                                  10002 => 'Incorrect number of parameters for Exception message; %s expected',
@ -37,7 +39,9 @@ class Exception extends \Exception {
                                  10401 => 'The Tokenizer has entered an invalid state',

                                  10501 => 'Form element expected, found %s',
-                                  10502 => 'Element, Document, or DOMDocumentFragment expected; found %s'];
+                                  10502 => 'Element, Document, or DOMDocumentFragment expected; found %s',
+
+                                  10601 => 'Non-empty Document supplied as argument for Parser'];

    public function __construct(int $code, ...$args) {
        if (!isset(static::$messages[$code])) {
--- a/lib/Parser.php
+++ b/lib/Parser.php
@ -53,10 +53,11 @@ class Parser {
    }

    public function __destruct() {
+        $this->treeBuilder->__destruct();
        static::$instance = null;
    }

-    public static function parse(string $data, bool $file = false) {
+    public static function parse(string $data, Document $document = null, bool $file = false) {
        // If parse() is called by parseFragment() then don't create an instance. It has
        // already been created.
        $c = __CLASS__;
@ -64,8 +65,14 @@ class Parser {
            static::$instance = new $c;
        }

-        if (is_null(static::$instance->DOM)) {
-            static::$instance->DOM = new DOM\Document();
+        if (is_null($document)) {
+            static::$instance->DOM = new Document();
+        } else {
+            if ($document->hasChildNodes()) {
+                throw new Exception(Exception::PARSER_NONEMPTY_DOCUMENT);
+            }
+
+            static::$instance->DOM = $document;
        }

        // Process the input stream.
@ -109,7 +116,7 @@ class Parser {
        $c = __CLASS__;
        static::$instance = new $c;

-        static::$instance->DOM = (is_null($context)) ? new DOM\Document() : $context->ownerDocument;
+        static::$instance->DOM = (is_null($context)) ? new Document() : $context->ownerDocument;
        static::$instance->DOMFragment = static::$instance->DOM->createDocumentFragment();

        // DEVIATION: The spec says to let the document be in quirks mode if the
--- a/lib/TreeBuilder.php
+++ b/lib/TreeBuilder.php
@ -76,7 +76,7 @@ class TreeBuilder {
    const QUIRKS_MODE_LIMITED = 2;


-    public function __construct(DOM\Document $dom, $formElement, bool $fragmentCase = false, $fragmentContext = null, OpenElementsStack $stack, Stack $templateInsertionModes, Tokenizer $tokenizer) {
+    public function __construct(Document $dom, $formElement, bool $fragmentCase = false, $fragmentContext = null, OpenElementsStack $stack, Stack $templateInsertionModes, Tokenizer $tokenizer) {
        // If the form element isn't an instance of DOMElement that has a node name of
        // "form" or null then there's a problem.
        if (!is_null($formElement) && !($formElement instanceof DOMElement && $formElement->nodeName === 'form')) {