How to separate HTTP header from binary content?

James_Bond_007

When I read a remote image file URL via fsockopen/fread, I get something like this:

HTTP 1.1/200 OK
... more headers...
Content-Type: image/gif

GIF89Adfsskdjfksd <binary content>

How can I make a string that contains only the binary part starting from GIF89A? substr doesn't work because of the zeros.

Installer

See the user's note by geoffrey at nevra dot net here.

Weedpacket

Also have a look at §4.1 of the HTTP Standard (http uses the same method of distinguishing headers and body as SMTP uses for email).

James_Bond_007

Thanks! I tried substr in a different way:

$nStartPos = strpos($sFullContent, "\r\n\r\n") + 4;
$sBinary = substr($sFullContent, $nStartPos);

This didn't seem to work.

Installer

If your PHP installation has curl enabled, you could do it like this:

$remote_img_url = 'http://www.example.com/image.gif';
$local_img_file = 'local_img.gif';
$content_type = 'Content-Type: image/gif';
$img = retrieve_remote($remote_img_url, $content_type);
$fp = fopen($local_img_file, 'wb');
fwrite($fp, $img);
fclose($fp);
echo '<img src="' . $local_img_file . '" />';

function retrieve_remote($url, $type) { 
   $ch = curl_init(); 
   curl_setopt($ch, CURLOPT_URL, $url); 
   curl_setopt($ch, CURLOPT_HTTPHEADER, array($type)); 
   curl_setopt($ch, CURLOPT_HEADER, 0); 
   curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1); 
   $data = curl_exec($ch); 
   curl_close($ch); 
   return $data; 
}

James_Bond_007

Unfortunately, I need to stick with standard PHP modules, because the code should be able to run almost anywhere.

Installer

This works for me:

$remote_img_url = 'http://www.example.com/image.gif';
$img = http_get($remote_img_url);
$local_img_file = 'local_img.gif';
$fp = fopen($local_img_file, 'wb');
fwrite($fp, $img);
fclose($fp);
echo '<img src="' . $local_img_file . '" />';

function http_get($url) 
{ 
    $url_stuff = parse_url($url); 
    if (isset($url_stuff['port'])) {
        $port = $url_stuff['port'];
    } else {
        $port = 80;
    } 
    $fp = fsockopen($url_stuff['host'], $port); 
    $query  = 'GET ' . $url_stuff['path'] . " HTTP/1.0\n"; 
    $query .= 'Host: ' . $url_stuff['host']; 
    $query .= "\n\n"; 
    fwrite($fp, $query); 
    $buffer = '';
    while ($tmp = fread($fp, 1024)) { 
       $buffer .= $tmp; 
    } 
    preg_match('/Content-Length: ([0-9]+)/', $buffer, $parts); 
    return substr($buffer, - $parts[1]);
}

James_Bond_007

The problem with this approach is that not all servers return Content-Length. For example, this server does not return Content-Length and does not know the difference between the HEAD and the GET command either. :-(

http://www.pbase.com/jamesbond007/image/33388967

Weedpacket

Maybe you should mention HEAD to your host 😉

I modified Installer's code and things look good for me. The thing to note is that the header is plain ASCII - no NULs nor empty lines except for the EOH. So you don't have to worry about those until after you've read the header.

This picks up from after the request is made:

$header = '';
   while ($tmp = fgets($fp, 1024)) { 
	 if($tmp=="\r\n") break;
    $header .= $tmp; 
    }
$content = '';
while($tmp=fread($fp,1024))
{
	$content.=$tmp;
}

James_Bond_007

Yep, I could tell them not to return the entire content for a HEAD request. 🙂

So this works for now for getting a page. At least until I try it on other servers.

   $fp = fsockopen($sHost, $nPort, $errno, $errstr, 30);

   if (!$fp) 
      {
      echo "$errstr ($errno)<br />\n";
      $sHeaders = NULL;
      $sContent = NULL;
      return;
      } 

   $out = "GET $sPath HTTP/1.0\r\n";
   $out .= "Host: $sHost\r\n";
   $out .= "Connection: Close\r\n\r\n";
   fwrite($fp, $out);

   $sHeaders = ''; 
   $sContent = ''; 
   $bHeaderCompleted = false; 

   while (!feof($fp))
   { 
      $tmp = fread($fp, READBUFSIZE);
      if (!$bHeaderCompleted)
         {
         $nHeaderEnd = strpos($tmp, "\r\n\r\n");
         if ($nHeaderEnd !== FALSE)
            {
            $sHeaders .= substr($tmp, 0, $nHeaderEnd);
            $sContent .= substr($tmp, $nHeaderEnd + 4);
            $bHeaderCompleted = TRUE;
            }
         else
            {
            $sHeaders .= $tmp;
            }
         }
      else
         {
         $sContent .= $tmp;
         } 
      }

   fclose ($fp); 
   }

Installer

Actually, the code I posted is not mine, except for the ternary statement translated to "if...else". Otherwise it's exactly the same code I linked to earlier.