How to decide whether to prepend www. to a URL in PHP?-CodePudding

I'm writing a PHP application where the user can enter a URL and some operations take place afterwards (further details not relevant to this question).

Requirement: If the user enters example.com, it should be converted to http://www.example.com.

The http:// part is straight-forward but am struggling with the rules that determine whether www. is prepended. Since the URL could be anything that might work in a web browser, it could be localhost or 192.168.0.1 for example. For these, clearly www. shouldn't be prepended.

So the exclusion list from above is: "If the host is localhost or looks like a v4 IP address, don't prepend". But expect there will be other cases that need to covered - could anyone advise - or suggest an alternative way of approaching this?

CodePudding user response：

You can validate the user input to IP and decide whether to concatenate the "www" or not. The user input can be "127.0.0.1", "127.0.0.1:8080","http://127.0.0.1:8080' or "http://exaple.com:8080".

$input = ("127.0.0.1:8080");
[$host,$port] = explode(":",trim($input,"http://"));
if(!empty($port)){
    $port=":".$port;
}
if (filter_var($host, FILTER_VALIDATE_IP)) {
    header("location:http://$host$port");
} else {
    header("location:www.$host$port");
}

CodePudding user response：

Here is my current attempt at doing this. It makes two passes because parse_url initially puts links without a scheme such as google.com or www.google.com into the "path" part rather than the "host" part.

function preprocessAbsoluteUrl($url, $firstRun = true) {
  $parts = parse_url($url);
  $scheme = isset($parts['scheme']) ? $parts['scheme'] : 'http';
  $user = isset($parts['user']) ? $parts['user'] : '';
  $pass = isset($parts['pass']) ? ":{$parts['pass']}" : '';
  $userpass = $user !== '' || $pass !== '' ? "{$user}{$pass}@" : '';
  $host = isset($parts['host'])
    ? (preg_match('/^(?:localhost|www\.|\d{1,3}\.\d{1,3}\d{1,3}\.\d{1,3})/i',
                  $parts['host'])
      ? $parts['host']
      : "www.{$parts['host']}")
    : '';
  $port = isset($parts['port']) ? ":{$parts['port']}" : '';
  $path = isset($parts['path']) ? rtrim($parts['path'], '/') : '';
  $query = isset($parts['query']) ? "?{$parts['query']}" : '';
  $fragment = isset($parts['fragment']) ? "#{$parts['fragment']}" : '';
  $url = "{$scheme}://{$userpass}{$host}{$port}{$path}{$query}{$fragment}";
  if ($firstRun) {
    $url = preprocessAbsoluteUrl($url, false);
  }
  return $url;
}

The relevant part is the setting of $host: This currently uses a regular expression to only prepend www. when it doesn't begin with www., localhost or look like an IP address. Open to improvement suggestions!