{
try
{
- $options = array('http'=>array('method'=>'GET','timeout' => $timeout)); // Force network timeout
+ $options = array('http'=>array('method'=>'GET','timeout' => $timeout, 'user_agent' => 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:23.0) Gecko/20100101 Firefox/23.0')); // Force network timeout
$context = stream_context_create($options);
$data=file_get_contents($url,false,$context,-1, 4000000); // We download at most 4 Mb from source.
if (!$data) { return array('HTTP Error',array(),''); }
{
list($status,$headers,$data) = getHTTP($url,4); // Short timeout to keep the application responsive.
// FIXME: Decode charset according to specified in either 1) HTTP response headers or 2) <head> in html
- if (strpos($status,'200 OK')!==false) $title=html_entity_decode(html_extract_title($data),ENT_QUOTES,'UTF-8');
-
+ if (strpos($status,'200 OK')!==false)
+ {
+ // Look for charset in html header.
+ preg_match('#<meta .*charset=.*>#Usi', $data, $meta);
+
+ // If found, extract encoding.
+ if (!empty($meta[0]))
+ {
+ // Get encoding specified in header.
+ preg_match('#charset="?(.*)"#si', $meta[0], $enc);
+ // If charset not found, use utf-8.
+ $html_charset = (!empty($enc[1])) ? strtolower($enc[1]) : 'utf-8';
+ }
+ else { $html_charset = 'utf-8'; }
+
+ // Extract title
+ $title = html_extract_title($data);
+ if (!empty($title))
+ {
+ // Re-encode title in utf-8 if necessary.
+ $title = ($html_charset == 'iso-8859-1') ? utf8_encode($title) : $title;
+ }
+ }
}
if ($url=='') $url='?'.smallHash($linkdate); // In case of empty URL, this is just a text (with a link that point to itself)
$link = array('linkdate'=>$linkdate,'title'=>$title,'url'=>$url,'description'=>$description,'tags'=>$tags,'private'=>0);
{
$attr=$m[1]; $value=$m[2];
if ($attr=='HREF') $link['url']=html_entity_decode($value,ENT_QUOTES,'UTF-8');
- elseif ($attr=='ADD_DATE') $raw_add_date=intval($value);
+ elseif ($attr=='ADD_DATE')
+ {
+ $raw_add_date=intval($value);
+ if ($raw_add_date>30000000000) $raw_add_date/=1000; //If larger than year 2920, then was likely stored in milliseconds instead of seconds
+ }
elseif ($attr=='PRIVATE') $link['private']=($value=='0'?0:1);
elseif ($attr=='TAGS') $link['tags']=html_entity_decode(str_replace(',',' ',$value),ENT_QUOTES,'UTF-8');
}
if (isset($_SERVER["QUERY_STRING"]) && startswith($_SERVER["QUERY_STRING"],'ws=')) { processWS(); exit; } // Webservices (for jQuery/jQueryUI)
if (!isset($_SESSION['LINKS_PER_PAGE'])) $_SESSION['LINKS_PER_PAGE']=$GLOBALS['config']['LINKS_PER_PAGE'];
renderPage();
-?>
\ No newline at end of file
+?>
projects / github / shaarli / Shaarli.git / blobdiff