Checking if a CDN file is available

Question

I have the following code working on a WordPress plugin. But there's no WP issue, the matter is:
can I trust @get_headers to perform most of times?

public function enqueue()
{
 $http = is_ssl() ? 'https:' : 'http:';
 $url = "$http//netdna.bootstrapcdn.com/font-awesome/3.2.0/css/font-awesome.min.csss";
 if( $this->get_http_response_code( $url ) ) {
 my_load_external_script();
 }
 else {
 my_load_local_script();
 }
}
/**
 * Problem with getting header, return false
 *
 * otherwise return if headers are 200 or not
 *
 * @param $url string
 * @return boolean
 */
private function get_http_response_code( $url ) 
{
 $headers = @get_headers( $url );
 if( !$headers )
 return false;
 return substr( $headers[0], 9, 3 ) === '200';
 }

The current requirement for running WordPress is PHP 5.2.4 or greater. As I understand it, there's no 100% guarantee, but for having a plugin running in the wild I'd prefer the less problematic path. I've checked the following Stack Overflow Q&A's, but haven't come to any conclusion:

• get_headers is throwing a warning instead of just returning false?

• Can I try/catch a warning?

• Reactgular: Hitting another server on a page request just to check if the CDN is available seems like bad design to me. I would do this check client side in javascript, and if the file is 404 then fallback to an alternate URL. Or make it a configured option of which URL to use.

Answer 1

In my experience no, using curl is much better an reliable. I use the following code in one of my projects, this always works (and allows redirects, if you don't want to set CURLOPT_MAXREDIRS to 0).

<?php
/**
 * Check if the URL is reachable.
 *
 * @param string $url
 * The URL to check.
 * @return this
 * @throws \LogicException
 */
public function checkReachability($url) {
 $ch = curl_init($url);
 curl_setopt_array($ch, [
 CURLOPT_AUTOREFERER => true,
 CURLOPT_CONNECTTIMEOUT => 5,
 CURLOPT_ENCODING => "",
 CURLOPT_FOLLOWLOCATION => true,
 CURLOPT_MAXREDIRS => 1,
 CURLOPT_NOBODY => true,
 CURLOPT_SSL_VERIFYHOST => false,
 CURLOPT_SSL_VERIFYPEER => false,
 CURLOPT_TIMEOUT => 5,
 // It's very important to let other webmasters know who's probing their servers.
 CURLOPT_USERAGENT => "Mozilla/5.0 (compatible; StackOverflow/0.0.1; +https://codereview.stackexchange.com/)",
 ]);
 curl_exec($ch);
 $code = curl_getinfo($ch, CURLINFO_HTTP_CODE);
 curl_close($ch);
 if ($code !== 200) {
 throw new LogicException("The URL doesn’t exists (more specifically isn’t reachable).");
 }
 return $this;
}
?>

Of course you have the problem that you don't always have curl available, a fallback would be.

<?php
// PHP 5.3+
stream_context_set_default(array("http" => array(
 "method" => "HEAD",
 "user_agent" => "Mozilla/5.0 (compatible; StackOverflow/0.0.1; +https://codereview.stackexchange.com/)",
)));
set_error_handler(function ($errno, $errstr, $errfile, $errline ) {
 throw new LogicException("The URL doesn’t exists (more specifically isn’t reachable).");
});
$headers = get_headers($url);
restore_error_handler();
// String comparisons should be made with ==
if (!isset($headers[0]) || substr($headers[0], 9, 3) != "200") {
 throw new LogicException("The URL doesn’t exists (more specifically isn’t reachable).");
}
?>

• brasofilo: Kool, I need to wait for new PC until I can test this, thx for the feedback.