From: Alexander Ebert <ebert@woltlab.com>
Date: Thu, 28 Jul 2016 10:18:04 +0000 (+0200)
Subject: Added output processor for Google AMP
X-Git-Tag: 3.0.0_Beta_1~946
X-Git-Url: https://git.stricted.de/?a=commitdiff_plain;h=cc8ed1659a87152d470fbf46e730679e9ed33d23;p=GitHub%2FWoltLab%2FWCF.git

Added output processor for Google AMP
---

diff --git a/wcfsetup/install/files/lib/system/html/output/AmpHtmlOutputProcessor.class.php b/wcfsetup/install/files/lib/system/html/output/AmpHtmlOutputProcessor.class.php
new file mode 100644
index 0000000000..215b0d44d1
--- /dev/null
+++ b/wcfsetup/install/files/lib/system/html/output/AmpHtmlOutputProcessor.class.php
@@ -0,0 +1,107 @@
+<?php
+namespace wcf\system\html\output;
+use wcf\util\DOMUtil;
+
+/**
+ * Processes stored HTML for final display and modifies/removes elements to
+ * match the Google AMP specifications.
+ * 
+ * See https://github.com/ampproject/amphtml/blob/master/spec/amp-html-format.md#html-tags
+ * 
+ * @author      Alexander Ebert
+ * @copyright   2001-2016 WoltLab GmbH
+ * @license     GNU Lesser General Public License <http://opensource.org/licenses/lgpl-license.php>
+ * @package     WoltLabSuite\Core\System\Html\Output
+ * @since       3.0
+ */
+class AmpHtmlOutputProcessor extends HtmlOutputProcessor {
+	/**
+	 * @inheritDoc
+	 */
+	public function process($html, $objectType, $objectID) {
+		parent::process($html, $objectType, $objectID);
+		
+		$document = $this->getHtmlOutputNodeProcessor()->getDocument();
+			
+		// remove tags and discarding content
+		$tags = [
+			// general
+			'base', 'frame', 'frameset', 'object', 'param', 'applet', 'embed',
+			
+			// forms
+			'input', 'textarea', 'select', 'option',
+			
+			// special
+			'style'
+		];
+		foreach ($tags as $tag) {
+			$elements = $document->getElementsByTagName($tag);
+			while ($elements->length) DOMUtil::removeNode($elements->item(0), true);
+		}
+		
+		// remove tags but keep child nodes
+		$tags = ['form'];
+		foreach ($tags as $tag) {
+			$elements = $document->getElementsByTagName($tag);
+			while ($elements->length) DOMUtil::removeNode($elements->item(0), false);
+		}
+		
+		// remove script tags unless the type is application/ld+json
+		$elements = $this->filterElements(
+			$document->getElementsByTagName('script'),
+			function ($element) {
+				/** @var \DOMElement $element */
+				return ($element->getAttribute('type') === 'application/ld+json');
+			}
+		);
+		foreach ($elements as $element) DOMUtil::removeNode($element);
+		
+		// replace tags
+		$tags = ['img', 'video', 'audio', 'iframe'];
+		foreach ($tags as $tag) {
+			$elements = $document->getElementsByTagName($tag);
+			while ($elements->length) {
+				/** @var \DOMElement $element */
+				$element = $elements->item(0);
+				if ($tag === 'img') {
+					$styles = $element->getAttribute('style');
+					if (preg_match('~\bheight:\s*(\d+)px\b~', $styles, $matches)) $element->setAttribute('height', $matches[1]);
+					if (preg_match('~\bwidth:\s*(\d+)px\b~', $styles, $matches)) $element->setAttribute('width', $matches[1]);
+					
+					if (!$element->getAttribute('height') || !$element->getAttribute('width')) {
+						DOMUtil::removeNode($element);
+						continue;
+					}
+					
+					$element->removeAttribute('style');
+				}
+				
+				$newElement = $element->ownerDocument->createElement('amp-' . $tag);
+				
+				// copy attributes
+				for ($i = 0, $length = $element->attributes->length; $i < $length; $i++) {
+					$attr = $element->attributes->item($i);
+					
+					$newElement->setAttribute($attr->localName, $attr->nodeValue);
+				}
+				
+				$element->parentNode->insertBefore($newElement, $element);
+				DOMUtil::removeNode($element);
+			}
+		}
+		
+		
+	}
+	
+	protected function filterElements(\DOMNodeList $elements, callable $callback) {
+		$badElements = [];
+		
+		foreach ($elements as $element) {
+			if ($callback($element) === false) {
+				$badElements[] = $element;
+			}
+		}
+		
+		return $badElements;
+	}
+}