From d8c39f2b0571b9734259b2f9dc218eed24412332 Mon Sep 17 00:00:00 2001
From: Christian Weiske <cweiske@cweiske.de>
Date: Thu, 11 Feb 2016 17:37:12 +0100
Subject: sanitize title better

---
 src/phinde/LinkExtractor/Html.php | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

(limited to 'src/phinde/LinkExtractor/Html.php')

diff --git a/src/phinde/LinkExtractor/Html.php b/src/phinde/LinkExtractor/Html.php
index 4acd19f..a6fa8ef 100644
--- a/src/phinde/LinkExtractor/Html.php
+++ b/src/phinde/LinkExtractor/Html.php
@@ -2,6 +2,7 @@
 namespace phinde\LinkExtractor;
 
 use phinde\LinkInfo;
+use phinde\Helper;
 
 class Html
 {
@@ -36,10 +37,10 @@ class Html
         $links = $dx->evaluate('//a');
         //FIXME: link rel, img, video
 
-        $alreadySeen = array();
+        $alreadySeen = array($url => true);
 
         foreach ($links as $link) {
-            $linkTitle = $link->textContent;
+            $linkTitle = Helper::sanitizeTitle($link->textContent);
             $href = '';
             foreach ($link->attributes as $attribute) {
                 if ($attribute->name == 'href') {
-- 
cgit v1.2.3