From 300b8bba9b229fd5dfbd496602195a173da46901 Mon Sep 17 00:00:00 2001
From: ORelio <oreliogitantispam.l0gin@spamgourmet.com>
Date: Fri, 23 Oct 2015 12:08:18 +0200
Subject: [PATCH 1/2] [Facebook] More cleaning and bypass link redirect

Remove a bunch of useless html node properties
Bypass all l.php facebook link redirections
Restore <u> and <i> basic formatting
---
 bridges/FacebookBridge.php | 40 ++++++++++++++++++++++++++++++++------
 1 file changed, 34 insertions(+), 6 deletions(-)

diff --git a/bridges/FacebookBridge.php b/bridges/FacebookBridge.php
index 109be9bb..3df13c34 100644
--- a/bridges/FacebookBridge.php
+++ b/bridges/FacebookBridge.php
@@ -3,7 +3,7 @@
 * @name Facebook
 * @homepage http://facebook.com/
 * @description Input a page title or a profile log. For a profile log, please insert the parameter as follow : myExamplePage/132621766841117
-* @update 05/09/2015
+* @update 23/10/2015
 * @maintainer teromene
 * @use1(u="username")
 */
@@ -13,6 +13,27 @@ class FacebookBridge extends BridgeAbstract{
 
 	public function collectData(array $param){
 
+		//Extract a string using start and end delimiters
+		function ExtractFromDelimiters($string, $start, $end) {
+			if (strpos($string, $start) !== false) {
+				$section_retrieved = substr($string, strpos($string, $start) + strlen($start));
+				$section_retrieved = substr($section_retrieved, 0, strpos($section_retrieved, $end));
+				return $section_retrieved;
+			} return false;
+		}
+
+		//Utility function for cleaning a Facebook link
+		$unescape_fb_link = function ($matches) {
+			if (is_array($matches) && count($matches) > 1) {
+				$link = $matches[1];
+				if (strpos($link, '/') === 0)
+					$link = 'https://facebook.com'.$link.'"';
+				if (strpos($link, 'facebook.com/l.php?u=') !== false)
+					$link = urldecode(ExtractFromDelimiters($link, 'facebook.com/l.php?u=', '&'));
+				return ' href="'.$link.'"';
+			}
+		};
+
 		$html = '';
 
 		if(isset($param['u'])) {
@@ -39,16 +60,23 @@ class FacebookBridge extends BridgeAbstract{
 
 				if($post->hasAttribute("data-time")) {
 
-					//Clean the content of the page and convert relative links into absolute links
+					//Retrieve post contents
 					$content = preg_replace('/(?i)><div class=\"clearfix([^>]+)>(.+?)div\ class=\"userContent\"/i', '', $post);
 					$content = preg_replace('/(?i)><div class=\"_59tj([^>]+)>(.+?)<\/div><\/div><a/i', '', $content);
 					$content = preg_replace('/(?i)><div class=\"_3dp([^>]+)>(.+?)div\ class=\"[^u]+userContent\"/i', '', $content);
 					$content = preg_replace('/(?i)><div class=\"_4l5([^>]+)>(.+?)<\/div>/i', '', $content);
-					$content = str_replace(' href="/', ' href="https://facebook.com/', $content);
-					$content = preg_replace('/ onmouseover=\"[^"]+\"/i', '', $content);
-					$content = preg_replace('/ onclick=\"[^"]+\"/i', '', $content);
+
+					//Remove html nodes, keep only img, links, basic formatting
+					$content = strip_tags($content,'<a><img><i><u>');
+
+					//Adapt link hrefs: convert relative links into absolute links and bypass external link redirection
+					$content = preg_replace_callback('/ href=\"([^"]+)\"/i', $unescape_fb_link, $content);
+
+					//Clean useless html tag properties and fix link closing tags
+					foreach (array('onmouseover', 'onclick', 'target', 'ajaxify', 'tabindex',
+						'class', 'style', 'data-[^=]*', 'aria-[^=]*', 'role', 'rel', 'id') as $property_name)
+							$content = preg_replace('/ '.$property_name.'=\"[^"]*\"/i', '', $content);
 					$content = preg_replace('/<\/a [^>]+>/i', '</a>', $content);
-					$content = strip_tags($content,'<a><img>');
 
 					//Retrieve date of the post
 					$date = $post->find("abbr")[0];

From c8ef31bac66b4ad710405886c4afdda7ae7eb0fb Mon Sep 17 00:00:00 2001
From: ORelio <oreliogitantispam.l0gin@spamgourmet.com>
Date: Fri, 23 Oct 2015 12:19:12 +0200
Subject: [PATCH 2/2] [NextInpact] Fix php notice message

For non-premium articles, a notice is generated because of trying to get
property from a non-object. Fixed that with a proper check using is_object
---
 bridges/NextInpactBridge.php | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/bridges/NextInpactBridge.php b/bridges/NextInpactBridge.php
index c91a4837..c070038a 100644
--- a/bridges/NextInpactBridge.php
+++ b/bridges/NextInpactBridge.php
@@ -8,7 +8,7 @@
 * @homepage http://www.nextinpact.com/
 * @description Returns the newest articles.
 * @maintainer qwertygc
-* @update 2015-09-05
+* @update 2015-10-23
 */
 class NextInpactBridge extends BridgeAbstract {
 
@@ -25,9 +25,9 @@ class NextInpactBridge extends BridgeAbstract {
 			$text = '<p><em>'.$html2->find('span.sub_title', 0)->innertext.'</em></p>'
 				.'<p><img src="'.$html2->find('div.container_main_image_article', 0)->find('img.dedicated',0)->src.'" alt="-" /></p>'
 				.'<div>'.$html2->find('div[itemprop=articleBody]', 0)->innertext.'</div>';
-			$premium_article = $html2->find('h2.title_reserve_article', 0)->innertext;
-			if (strlen($premium_article) > 0)
-				$text = $text.'<p><em>'.$premium_article.'</em></p>';
+			$premium_article = $html2->find('h2.title_reserve_article', 0);
+			if (is_object($premium_article))
+				$text = $text.'<p><em>'.$premium_article->innertext.'</em></p>';
 			return $text;
 		}