Merge pull request #1055 from annando/master

[friendica.git] / mod / parse_url.php

diff --git a/mod/parse_url.php b/mod/parse_url.php
index a0608f636340cfe8e88b41f5c1bae123204974f6..7ab71a2fc520d7f7ae68a07711a7cc35fe947633 100644 (file)
--- a/mod/parse_url.php
+++ b/mod/parse_url.php
@@ -50,9 +50,15 @@ function completeurl($url, $scheme) {
         return($complete);
 }
 
-function parseurl_getsiteinfo($url, $no_guessing = false) {
+function parseurl_getsiteinfo($url, $no_guessing = false, $do_oembed = true, $count = 1) {
+
        $siteinfo = array();
 
+       if ($count > 10) {
+               logger("parseurl_getsiteinfo: Endless loop detected for ".$url, LOGGER_DEBUG);
+               return($siteinfo);
+       }
+
        $url = trim($url, "'");
        $url = trim($url, '"');
        $siteinfo["url"] = $url;
@@ -65,7 +71,8 @@ function parseurl_getsiteinfo($url, $no_guessing = false) {
        curl_setopt($ch, CURLOPT_TIMEOUT, 3);
        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
        //curl_setopt($ch, CURLOPT_FOLLOWLOCATION, 1);
-       curl_setopt($ch,CURLOPT_USERAGENT,' Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:24.0) Gecko/20100101 Firefox/24.0');
+       //curl_setopt($ch,CURLOPT_USERAGENT,' Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:24.0) Gecko/20100101 Firefox/24.0');
+       curl_setopt($ch,CURLOPT_USERAGENT, "Mozilla/5.0 (compatible; Friendica)");
 
        $header = curl_exec($ch);
        $curl_info = @curl_getinfo($ch);
@@ -75,18 +82,20 @@ function parseurl_getsiteinfo($url, $no_guessing = false) {
        if ((($curl_info['http_code'] == "301") OR ($curl_info['http_code'] == "302") OR ($curl_info['http_code'] == "303") OR ($curl_info['http_code'] == "307"))
                AND (($curl_info['redirect_url'] != "") OR ($curl_info['location'] != ""))) {
                if ($curl_info['redirect_url'] != "")
-                       $siteinfo = parseurl_getsiteinfo($curl_info['redirect_url']);
+                       $siteinfo = parseurl_getsiteinfo($curl_info['redirect_url'], $no_guessing, $do_oembed, ++$count);
                else
-                       $siteinfo = parseurl_getsiteinfo($curl_info['location']);
+                       $siteinfo = parseurl_getsiteinfo($curl_info['location'], $no_guessing, $do_oembed, ++$count);
                return($siteinfo);
        }
 
-       require_once("include/oembed.php");
+       if ($do_oembed) {
+               require_once("include/oembed.php");
 
-       $oembed_data = oembed_fetch_url($url);
+               $oembed_data = oembed_fetch_url($url);
 
-       if ($oembed_data->type != "error")
-               $siteinfo["type"] = $oembed_data->type;
+               if ($oembed_data->type != "error")
+                       $siteinfo["type"] = $oembed_data->type;
+       }
 
        // Fetch the first mentioned charset. Can be in body or header
        $charset = "";
@@ -139,7 +148,7 @@ function parseurl_getsiteinfo($url, $no_guessing = false) {
                                         $content = substr($value, 4);
                         }
                         if ($content != "") {
-                                $siteinfo = parseurl_getsiteinfo($content);
+                                $siteinfo = parseurl_getsiteinfo($content, $no_guessing, $do_oembed, ++$count);
                                 return($siteinfo);
                         }
                 }
@@ -172,7 +181,7 @@ function parseurl_getsiteinfo($url, $no_guessing = false) {
                                        $siteinfo["image"] = $attr["content"];
                                        break;
                                case "twitter:card":
-                                       if ($siteinfo["type"] == "")
+                                       if (($siteinfo["type"] == "") OR ($attr["content"] == "photo"))
                                                $siteinfo["type"] = $attr["content"];
                                        break;
                                case "twitter:description":
@@ -216,7 +225,7 @@ function parseurl_getsiteinfo($url, $no_guessing = false) {
                        }
        }
 
-       if ($oembed_data->type == "link") {
+       if (isset($oembed_data) AND ($oembed_data->type == "link") AND ($siteinfo["type"] != "photo")) {
                if (isset($oembed_data->title) AND (trim($oembed_data->title) != ""))
                        $siteinfo["title"] = $oembed_data->title;
                if (isset($oembed_data->description) AND (trim($oembed_data->description) != ""))
@@ -440,3 +449,4 @@ function parse_url_content(&$a) {
 
        killme();
 }
+?>