Image headers returning text/html

Question

I'm attempting to retrieve images from a web page, and it has been working well so far, except one of the sites I am looking at is serving images as Content-Type: text/html , causing my script to reject it as not a real image.

This is the code snippet I am using to determine content-type:

$accepted_mime = array('image/gif', 'image/jpeg', 'image/jpg', 'image/png');    
$headers = get_headers($image);

// Find the Content-Type header
$num_headers = sizeOf($headers);
for($x=0;$x<$num_headers;$x++) {
    preg_match('/^Content-Type: (.+)$/', $headers[$x], $mime_type);
    if (isset($mime_type[1]) && in_array($mime_type[1], $accepted_mime)) {
        return true;
    }
}

For sites I've tried, they return properly (results such as image/gif , image/png , etc), but mpaa.org seems to serve their images with type text/html . Is this normal?

I added a print_r to see the header array returned by get_headers`:

Array
(
    [0] => http://www.mpaa.org/templates/images/header_mpaa_logo.gif
    [1] => Array
        (
            [0] => HTTP/1.1 200 OK
            [1] => Server: nginx/1.2.0
            [2] => Date: Sat, 17 Nov 2012 17:19:06 GMT
            [3] => Content-Type: text/html
            [4] => Connection: close
            [5] => P3P: CP="NON DSP COR ADMa OUR IND UNI COM NAV INT"
            [6] => Cache-Control: no-cache, no-store, must-revalidate
            [7] => Pragma: no-cache
        )

)

I could easily add text/html to my list of accepted content-types, but that's definitely not the ideal solution ;) Does anyone know why mpaa.org serves their images with this Content-Type ? Is it regular practice to do so (perhaps with legacy websites/servers)?

Thanks :)

Answer 1

The wonderful MPAA is using user-agent sniffing or checking cookies to determine if your browser supports JavaScript. Since you are not specifying a user-agent string or sending cookies, they assume you don't have JavaScript and return a page saying that, instead of the original image.

If you load this with a browser, you'll note that you do get image/gif , and the image you are after: http://www.mpaa.org/templates/images/header_mpaa_logo.gif

If you make that same request with cURL or Fiddler, or some other oddball user-agent string:

This site requires JavaScript and Cookies to be enabled. Please change your browser settings or upgrade your browser.

Answer 2

Dont rely on headers. They can be changed easily and as you encounter now, are not reliable.

I would do it like this:

• Download the image
• Check if the image is an image (by using getimagesize or something like that)