NYCPHP Meetup

[Phillip Powell]

What is the most standardized method of utilizing the CURL functions in 
PHP (version 4.3.2) to be able to retrieve the contents of a remote URL 
that happens to be dependent upon $_SESSION for its content display?

I've tried the following class methods for display and I have most 
everything working until I get to a URL that requires $_SESSION:

[PHP]
class Timer extends View {

   function Timer() {}

    function &setCookieCurlSetOpt(&$ch) {        // STATIC VOID METHOD
        if ($this->cookieName) {
         $qs = '&' . $this->cookieName . '=';
         if (is_array($_COOKIE[$this->cookieName]) || 
is_object($_COOKIE[$this->cookieName])) {
          $qs .= urlencode(serialize($_COOKIE[$this->cookieName]));
         } else {
          $qs .= urlencode($_COOKIE[$this->cookieName]);
         }
        } elseif (@sizeof(array_values($_COOKIE)) > 0) {
         foreach ($_COOKIE as $key => $val) if (is_array($val) || 
is_object($val)) $qs .= "&$key=" . urlencode(serialize($val)); else $qs 
.= "&$key=" . urlencode($val);
        }
        if ($qs) curl_setopt($ch, CURLOPT_COOKIE, substr($qs, 1, 
strlen($qs)));
    }

      function &displayHTML($url = '') {            // STATIC HTML 
STRING METHOD
        global $projectFolderName, $PHPSESSID;
        if (is_object($this) && !$this->getURL() && $url) 
$this->setURL($url);
        if (is_object($this)) $url = $this->getURL();
        if ($url && ini_get('allow_url_fopen')) {
         $url = $this->configureURL($url);
         // grab URL and pass it to the browser
         $ch = curl_init();
         $this->setCookieCurlSetOpt($ch);
         //$this->setPOSTCurlSetOpt($ch);
         curl_setopt($ch, CURLOPT_HEADER, 0);
         curl_setopt($ch, CURLOPT_URL, $url);
         $this->setTime($this->startTime);    // SET START TIME
         ob_start();
          curl_exec($ch);
          $html = ob_get_contents();
         ob_end_clean();
         $timer = $this->setTime($timer);
         // close curl resource, and free up system resources
         curl_close($ch);
         $html = preg_replace('/(.*<body[^>]*>)/i', '$1<p><b>' . ($timer 
- $this->startTime) . " seconds to run URL: \"$url\"</b>", $html);
        }
        return $html;
    }

}
[/PHP]

utilizing this class in a script: timer.php:

<?php

  session_cache_limiter();
  session_start();
  echo '<P>My session from timer.php is: '; print_r($_SESSION);
...

?>

I see my $_SESSION variable.

Now, if timer.php instantiates a Timer object that uses the URL 
parameter of "http://www.myphpwebsite.com/index.php" it will scrape that 
site.  The site's script: index.php, also has the following:

<?php

  session_cache_limiter();
  session_start();
  echo '<P>My session from index.php is: '; print_r($_SESSION);
...
?>

Now here's the kicker:  If I call index.php directly, everything's fine, 
$_SESSION exists, all is fine.  If I call index.php as a scrape via 
timer.php, then timer.php displays $_SESSION but index.php's scrape 
shows nothing, just an empty array instead of $_SESSION with its contents:

[timer.php]
My session from timer.php is:  Array([lots of stuff])

My session from index.php is: Array()
[/timer.php]

Is there something about autoglobals and CURL in general, or is there 
something specific about $_SESSION and CURL, or perhaps session_start() 
and CURL?  I am not sure what I can do about this dilemma as it means my 
Timer object will not accurately show how long it took for index.php to 
evaluate and download since the returned content is wrong due to the 
seeming nonexistence of $_SESSION all of a sudden ($_SESSION is not 
changed at any time during the run of timer.php)

I hope this is detailed enough to provide a clear picture of exactly 
what is going on and will allow for someone to give me further insight 
into the workings of CURL with $_SESSION or autoglobals in general.

Thanx
Phil
 
-- 

---------------------------------------------------------------------------------
Phil Powell
Multimedia Programmer
BPX Technologies, Inc.
#: (703) 709-7218 x107 
Fax: (703) 709-7219